19results about How to "Increase contribution" patented technology

The invention discloses a Co3O4-Cu2-xS composite photo-thermal catalyst as well as a preparation method and application thereof. The composite photo-thermal catalyst is prepared by taking a surfactant as a bridge and compounding Co3O4 and Cu2-xS. The composite photo-thermal catalyst has a reduced band gap width, can effectively promote migration and separation of photon-generated carriers, enhances the absorption efficiency of visible light-near infrared light, and shows excellent photo-thermal conversion performance. The composite photo-thermal catalyst can be used for constructing a Co3O4-Cu2-xS + PMS + NIR catalytic system, and peroxymonosulfate is activated under the irradiation of near-infrared light so as to degrade antibiotics in water. Experiments prove that the Co3O4-Cu2-xS + PMS + NIR system disclosed by the invention can achieve the degradation rate of 99% on levofloxacin within 20 minutes, and the generation efficiency of active species is remarkably improved through the synergistic effect of a photothermal effect and interface activation. The catalyst has broad-spectrum degradation activity on various antibiotics, keeps efficient and stable catalytic performance in a wide pH range, and shows excellent ion interference resistance and adaptability to various water bodies.

According to the metabolism-oriented clustering-based tumor-associated macrophage metabolism typing method provided by the invention, the single-cell RNA sequencing data set for analysis is collected and arranged, and then the tumor-associated macrophage population is identified and extracted from the single-cell RNA sequencing data set, so that the technical deviation is eliminated; a core metabolism gene set is constructed based on a tumor-associated macrophage population, then weight distribution and optimization of core metabolism genes are carried out, contribution of the metabolism genes in clustering analysis is enhanced, metabolism-oriented clustering is carried out based on the optimized weight, and a metabolism typing result is generated. Therefore, enhancement of metabolic gene expression characteristics and more accurate distinguishing of TAMs subgroups are achieved, typing results are analyzed and verified through GSEA enrichment analysis or GO enrichment analysis, typing categories are output, challenges confronted when high-dimensional single-cell data are processed through a traditional method are effectively overcome, the method is not only suitable for TAMs, but also can be expanded to other tumor-related cell types, and the method has a good application prospect. And the blank in the prior art is filled.

The invention discloses an intelligent detection method and system for abnormal SQL statement judgment, and the method comprises the steps: designing a CNN-LSTM dual-channel architecture and a cross-modal attention fusion layer through the introduction of a combined design of an abstract syntax tree, byte pair coding and a recurrent neural network, enabling the two networks to capture local and global features respectively, and carrying out the recognition of abnormal SQL statements. The feature weight is dynamically adjusted through an attention mechanism, deep interactive fusion is realized, a heterogeneous base learner cluster is constructed, a meta learner is constructed by extracting multi-dimensional scene features and introducing a multi-layer perceptron, the base learner weight adaptive to the scene is dynamically generated, and meanwhile, a judgment threshold value is adjusted in combination with actual requirements, so that scene self-adaptive accurate decision is realized; and finally, abnormal SQL detection is achieved. According to the method, the feature fusion defect is solved through a cross-modal dynamic attention mechanism, scene adaptive optimization is realized through a meta-learning-driven dynamic decision framework, and the diversity of the model is enhanced through cooperation of multi-scale embedding and differentiated training strategies.

The application discloses a social media-based interpretable dynamic graph network benefit prediction model training method, relates to the technical field of text analysis and market prediction, and obtains market data and related social media data of a financial asset to divide the data by weeks; for the social media data of each week, a topic is taken as a node, and similarity between topics is taken as an edge weight, a topic correlation graph of each week is constructed to obtain weekly graph data; a prediction model performs time sequence updating and market prediction according to the weekly graph data; a text time sequence memory unit obtains a current week topic memory vector according to a current week topic text embedding vector and in combination with a last week topic memory vector; a multi-layer graph attention network captures the correlation features between nodes through an attention mechanism and adopts attention weights for weighted summation to obtain a current week global graph embedding; and a classifier outputs a next week price change prediction result. The application solves the problem that the prior art cannot effectively capture the time sequence evolution of topic sentiment and the correlation between topics.

The application claims a kind of dual-channel and attention mechanism aspect-level multi-modal sentiment analysis method, which is based on neural network, the sentiment information contained in image feature is extracted by aspect word feature and sentence feature joint attention mechanism multi-scale, and GCN network is introduced into aspect-level multi-modal sentiment analysis task, greatly improve the feature extraction and interactive fusion ability of model.In the present application, the pre-training encoder is used to extract aspect word, sentence feature and image feature in the feature extraction layer.In the attention mechanism layer, after the bidirectional fusion of aspect word and sentence feature, the final aspect word feature and sentence feature representation are obtained.Image feature establishes image feature extraction network through channel attention mechanism and spatial attention mechanism.Finally, the interactive fusion features of each modality are dynamically extracted by GCN module.In the experiment, the performance indicators of aspect-level multi-modal sentiment analysis based on attention mechanism are improved on the data set.

The invention provides a loss prediction method and device based on local performance estimation, equipment and a medium, and relates to the field of artificial intelligence, and the method comprises the steps: obtaining feature data of a to-be-predicted sample, inputting the feature data into each base model, and obtaining the original prediction output of each base model; obtaining a neighbor sample set of the to-be-predicted sample in the historical verification set, and calculating a local performance index of the base model in the neighbor sample set for each base model; and calculating a dynamic adjustment coefficient of each base model according to the local performance index of each base model and the global performance index of each base model, generating a fusion weight for the to-be-predicted sample in combination with the global weight and the dynamic adjustment coefficient of the base model, and performing weighted fusion on the original prediction output by using the dynamic fusion weight to obtain a loss prediction result. The method can adapt to telecommunication user groups with high data heterogeneity, and the loss prediction precision of individual users is improved.

This invention discloses a single-target tracking method and system based on channel attention and spatiotemporal awareness. The method includes the following steps: inputting a template image and a search region image; extracting hierarchical features from both using two weight-shared RepVGG backbone networks; performing self-attention and cross-attention feature fusion on the two feature maps using a fusion model based on channel attention mechanism; completing target tracking prediction through classification branches, bounding box regression branches, and centrality regression branches; updating the template in real time through threshold judgment; adding the temporal information of subsequent frames to the fusion network to achieve complete target tracking; and demonstrating tracking performance. The system includes: a feature extraction module, a feature fusion module, a tracking prediction module, and a performance display module. This invention allows the template features to adaptively enhance their completeness with spatiotemporal changes, thereby improving the robustness of tracking.

The invention provides a preparation method of mulberry and Weishi wine and the obtained mulberry and Weishi wine, the preparation method of the mulberry and Weishi wine comprises the following steps: S1, adding hops into mulberry juice, and sterilizing to obtain a to-be-fermented liquid; s2, main fermentation: inoculating the activated and cultured saccharomyces cerevisiae BYBC 2.21108 into the to-be-fermented liquid for liquid fermentation; the preservation number of the saccharomyces cerevisiae BYBC 2.21108 is CGMCC (China General Microbiological Culture Collection Center) And S3, ageing to obtain the mulberry-Wei Tai wine. The aroma of the mulberry-Wei Di wine is dominated by sulfides, the mulberry-Wei Di wine is more unique, and the fermentation degree and the quality are remarkably improved. In the aspect of contribution of aroma main components, the fruit-flavor beer has the due beer flavor of beer and the aroma of fruits, and also has the characteristics of low alcoholic strength, rich organic acid and oxidation resistance, so that the fruit-flavor beer has great development and utilization values.

The application provides a distance measurement method and an electronic device, and relates to the field of visual perception. The method comprises the following steps: acquiring distance prediction values and uncertainties of a target object in a to-be-measured image determined by at least two distance measurement algorithms; wherein the uncertainty is set according to a prediction confidence determined by the distance measurement algorithm or is set according to at least two distance prediction values determined by the distance measurement algorithm; determining a fusion weight corresponding to each distance prediction value according to the uncertainty; and fusing each distance prediction value according to each fusion weight to obtain a distance measurement value. The distance prediction values of multiple distance measurement algorithms can be fused according to the uncertainty, so that the advantages of different ranging algorithms are fused for distance measurement, and the stability and reliability of monocular visual ranging can be improved.

The invention discloses a breeding environment pig cough identification method, device and equipment based on multi-mode audio, and a medium. The method comprises the following steps: acquiring a plurality of pig sound audio clips of a pig breeding site; inputting the plurality of pig sound audio clips into the trained pig cough sound recognition model to obtain a pig cough recognition result; the system for training the pig cough sound recognition model comprises a multi-mode spectrogram generation module which is used for generating a multi-mode audio spectrogram corresponding to each pig sound audio clip; the feature initialization module is used for converting the plurality of multi-mode audio spectrograms to obtain a first image feature corresponding to each mode; the feature extractor is used for performing feature extraction on the first image feature corresponding to each mode to obtain a second image feature corresponding to each mode; and the feature dynamic fusion output module is used for performing feature dynamic fusion on the second image features corresponding to each mode, and obtaining a pig cough recognition result through a classification layer based on the fused features.

The invention discloses a solid waste monitoring and data management system based on artificial intelligence, and relates to the technical field of water environment monitoring, the system is composed of an image acquisition module, a data preprocessing module, an encoder feature extraction module, an up-sampling fusion module, a detection head module and a data management module, and the system is based on an RW-YOLOv11 architecture. A C3K2Sc feature extraction unit is introduced into an encoder, and floating garbage edge and texture expression is enhanced through space attention and a dynamic channel reconstruction mechanism; a SurfCAU water surface content awareness enhanced up-sampling module is adopted in the neck network, detail compensation of low-resolution features is achieved, and the multi-scale feature fusion quality is improved; a SurfMSDFHead multi-branch structure is adopted in the detection head, multi-scale feature fusion is realized through dynamic weight, and the positioning precision of overlapped garbage is improved by using a Focaler-IOU interval weighting strategy. The system can realize real-time detection, classification and positioning of the floating garbage in the river channel, and has the advantages of high detection precision, good lightweight degree, strong adaptability to complex water surface scenes and the like.

The invention relates to the technical field of artificial intelligence and voice processing, can be applied to the field of intelligent medical treatment and finance, and discloses an audio forgery detection method and device, computer equipment and a storage medium, and the method comprises the steps: carrying out the preprocessing of a to-be-detected audio signal, and extracting the frequency domain features and time domain features of the preprocessed audio signal, obtaining a time domain feature sequence and a frequency domain feature sequence; according to the time-domain feature sequence and the frequency-domain feature sequence, establishing correlation between self-modals and cross-modals of the time-domain features and the frequency-domain features in a graph neural network mode by using an attention mechanism, and obtaining a fusion feature map of fusion of the time-domain features and the frequency-domain features; calculating a global feature of the fusion feature spectrum according to each node feature in the fusion feature spectrum; and adopting a classifier to carry out audio forging detection on the global features of the fusion feature spectrum.

This invention relates to a method for predicting the performance of marine turbocharger compressors via a performance map. It addresses the problems of existing methods, such as multiple constraint conflicts, local distortion, and the inability to simultaneously satisfy multiple physical consistency requirements, and belongs to the field of marine power. The invention includes: acquiring sparse data point sets containing speed, flow rate, pressure ratio, and efficiency at multiple speeds, and grouping them by speed. For a target speed, two adjacent subsets of speed lines are selected, and feature point data for the target speed are constructed using linear interpolation. The similarity between each known speed line and the target speed is calculated, including speed difference, peak position flow rate difference, peak amplitude difference, surge, stagnation endpoint flow rate difference, local slope difference, and curvature difference. Adaptive weights are generated based on the similarity, and the pressure ratio and efficiency curves of all known speed lines are weighted for prediction to obtain the target speed prediction curve. Joint constraint solving is performed to construct an objective function and apply multiple types of constraints to form a prediction result that satisfies all constraints, ultimately generating a complete compressor performance map.

This invention relates to the field of sound recognition technology for flying animals, and discloses a method for intelligent sound recognition of flying animals based on feature fusion. The method includes acquiring the sound signal of the target flying animal and performing time-frequency transformation to obtain a time-frequency feature matrix. Local feature vectors are extracted from the matrix, their gradient fields are calculated, local extrema are located, and their directional connectivity is encoded to form a local topological structure. High-energy extrema are selected as key local features. The Mel-frequency cepstral coefficient sequence of the signal is extracted, the statistical characteristics of each Mel-frequency band are calculated, and the coefficients corresponding to high-variance frequency bands are selected as key acoustic features. After fusing the two types of key features, the result is input into a pre-trained convolutional neural network to complete category recognition. This method enhances the feature discrimination ability by focusing on the essential structure and significant change information of sound, thereby improving recognition accuracy and robustness.

The invention provides a railway safety monitoring method and device based on multi-modal fusion, equipment and a medium, and belongs to the technical field of railway safety monitoring, and the method comprises the steps: obtaining multi-modal monitoring data of a railway in a preset time period, extracting a multi-modal alignment feature in semantic alignment with the railway from the multi-modal monitoring data, and storing the multi-modal alignment feature in the preset time period; and performing deep fusion on the multi-modal alignment features by adopting a cross-modal attention fusion mechanism to obtain fusion features, generating a railway fault semantic representation vector through semantic matching and feature mapping based on the fusion features and a pre-constructed railway multi-modal knowledge base, and performing feature fusion on the railway fault semantic representation vector. And inputting the railway fault semantic representation vector into a trained railway multi-mode safety monitoring model, and outputting the safety state probability of the railway in a preset time period. According to the railway safety monitoring method and device based on multi-modal fusion, the equipment and the medium provided by the invention, the recognition effect of railway safety monitoring can be improved.

This invention provides an optimization method for a temporally dependent data prediction model based on agent trajectory feedback, applicable to tasks such as webpage evidence text collection, open encyclopedia question answering, enterprise knowledge base question answering, or retrieval enhancement generation. The method collects multi-round execution trajectories formed by the agent's interaction with the retrieval system during task execution. Initial supervised samples are constructed based on the candidate document set returned by the search action and subsequent browsing actions. Positive samples are filtered using post-browsing inference text and an inference-evidence consistency judgment model. Relevance strength weights are estimated based on inference length, evidence citation count, number of fact entries, and changes in subsequent actions. The retrieval model is trained using temporally gated coding and a weighted contrastive learning objective function. The optimized retrieval model is then deployed back to the agent system for continuous closed-loop updates, improving the execution efficiency of webpage evidence collection, evidence retrieval, and complex question answering tasks.

The invention discloses a method for improving rolling uniformity of a multi-core superconducting strip. The method specifically comprises the steps that a plurality of core materials are subjected to tubing treatment, then the composite pipe subjected to tubing treatment is subjected to drawing, rolling and heat treatment, the core materials are regular hexagon core materials, and the core materials are single-core core materials or core materials obtained through tubing treatment. According to the invention, the regular hexagon single core (or multi-core) is used for secondary (or multiple) tubing, so that the deformation of the multi-core round wire in the drawing process is changed, the section of the superconducting core is changed from an original irregular shape to an approximately circular shape, and the stress condition of the single superconducting core in the rolling process is closer to that of a single-core strip; and finally, the superconducting core with uniform density and texture degree is obtained. In addition, geometric stability can be enhanced, and uncertain deformation is avoided.

The invention relates to the technical field of data analysis, can be applied to business system platforms of financial science and technology, medical health and the like, and discloses a key frame feature extraction method, device, equipment and medium for multi-modal data, and the method comprises the following steps: carrying out differentiable neural architecture search on a plurality of multi-modal original data streams to obtain multi-modal undetermined features; carrying out importance analysis on the multi-modal undetermined features to obtain importance scores, screening out an effective modal feature set from the multi-modal undetermined features, carrying out dynamic sparse connection on the effective modal feature set to obtain fused modal features, carrying out time sequence attention distillation on the fused modal features to obtain frame-level attention weights, and carrying out time sequence attention distillation on the fused modal features to obtain frame-level attention weights; and carrying out frame-level key analysis on the multi-modal original data stream to obtain a key frame of the multi-modal original data stream. According to the method, the unified modeling capability of the multi-modal time sequence is improved, dynamic adjustment of the modal weight and redundant information compression are realized, and the overall processing efficiency is remarkably improved while the calculation burden is reduced.

The present application relates to a kind of tomography methods for fast small-scale ionospheric disturbance event, based on the disturbance event occurrence time period of target ionosphere in the disturbance event occurrence space above, and with the various navigation satellites to be analyzed of target layout area in the space link of GNSS receiver, according to the GNSS receiver site selection position coordinate range in target time period, analyze the space intersection of the local motion coordinate track of each navigation satellite to be analyzed and the space link between GNSS receiver in disturbance event occurrence space, and with maximum space intersection as target, determine the optimal site selection coordinate of GNSS receiver, then according to the space link communication between GNSS receiver and navigation satellite to be analyzed, complete the tomography of disturbance event in target ionosphere, to improve effective observation time, the contribution degree of GNSS receiver monitoring data in disturbance event occurrence space to ionospheric tomography, solve the problem that the number of receiver is limited in the application of sparse distribution in fast small-scale ionospheric disturbance event tomography, effective monitoring time is short, coverage area is not enough etc., can effectively improve the ability of ionospheric tomography.