44results about How to "Highlight the importance" patented technology

The invention discloses an LIB SOC estimation method based on intelligent adaptive extended Kalman filtering, belongs to the field of power battery management. The LIB SOC estimation method comprisesthe following specific steps of establishing a state equation and a measurement equation according to a RC equivalent circuit, and discretizing to obtain an LIB state space model; constructing a polynomial open-circuit voltage model according to open-circuit voltage and SOC test data and obtaining model parameters; substituting an open-circuit voltage model into a measurement equation, obtaining current and terminal voltage through HPPC test, aiming at minimizing the error between the actually measured terminal voltage and an estimated value, and obtaining equivalent circuit model parameters based on a genetic algorithm; estimating the SOC of the LIB based on intelligent adaptive extended Kalman filtering, wherein the method adopts a dynamic window length innovation sequence to estimate noise and can adapt to the measurement noise change caused by dynamic working conditions; and the comparison result shows that the method has higher estimation precision and better robustness.

The invention discloses a refrigerating machine room energy-saving optimization method and system based on chilled water and cooling water variable flow decoupling. According to the method, on the basis of an original water chilling unit MP model, a cooling water inlet temperature modeling method based on a GRNN is applied to improve the model. Through equipment modeling, system simulation, orthogonal test and regression analysis, when the cold load of a building is determined, an orthogonal test method suitable for variable-flow decoupling of chilled water and cooling water in a refrigeratingmachine room is used; parameter values of a chilled water flow, a cooling water flow, a chilled water supply temperature and a cooling water inlet temperature which enable the total energy consumption of the refrigerating machine room to be minimum under the condition are determined, the energy consumption of each piece of equipment and system before and after optimization is compared, and each piece of equipment is adjusted correspondingly; therefore, theoretical basis and technical guidance are provided for energy-saving transformation of a water chilling unit and a refrigerating machine room.

The invention provides an image sparse representation-based robust human face identification method. The method comprises the steps of firstly performing multi-direction and multi-scale Gabor transformation on a human face image; secondly fusing Gabor features of different scales in the same direction to obtain multi-direction feature graphs; thirdly extracting Gist features from the multi-direction feature graphs and endowing the Gist features with different weights; fourthly cascading the weighted Gist features of the multi-direction feature graphs to form a human face image eigenvector; andfinally performing classification by utilizing sparse representation to realize human face identification. The method can solve the technical problem of relatively low identification rate of a humanface identification system caused by variable factors such as illumination, poses, expressions and the like; the features extracted by the method are good in representation capability; and a better human face classification effect is achieved.

The invention provides a visual question-answering method and system based on semantic alignment and a storage medium, and relates to the technical field of visual question-answering. According to theembodiment of the invention, the method comprises the steps: firstly obtaining and preprocessing a data set, extracting original image features and target position features according to an original image, generating an image description statement according to the target position features, obtaining an image description word, question features and image description statement features, and carryingout the semantic alignment of the original image features and the image description word; and obtaining a first image feature, obtaining a second image feature according to the original image featureand the image description statement feature, obtaining a third image feature according to the original image feature and the question feature, fusing the three image features, the image description statement feature and the question feature to obtain a comprehensive feature, and predicting a final answer result. Therefore, the importance of the image information is highlighted, the information involved in the feature fusion process is perfected, and the finally generated answer result is more accurate.

A similar case retrieval method based on a multidimensional correlation function includes the steps of firstly, improving the lateral spacing through a variable slope method, and establishing a multidimensional lateral spacing model based on the improved lateral spacing; secondly, establishing the multidimensional correlation function with a good dimensionality reduction effect under the relatively-complex multidimensional lateral spacing calculation; thirdly, putting forward a three-stage low-carbon product retrieval method based on the multidimensional correlation function. By means of the similar case retrieval method based on the multidimensional correlation function, the data dimensionality calculation can be reduced, the retrieval efficiency can be improved, and the more accurate and reliable result can be obtained.

The invention relates to the technical field of artificial intelligence, and particularly discloses a text abstract generation method, device and equipment and a storage medium; the method comprises the following steps: obtaining a to-be-processed text, and inputting the to-be-processed text into a selector in an abstract generation model for keyword extraction to obtain a keyword sequence; inputting the keyword sequence into a first encoder in an abstract generation model for encoding processing to obtain a first vector; inputting the to-be-processed text into a second encoder in the abstract generation model for encoding processing to obtain a second vector; obtaining a hidden vector according to the first vector and the second vector; and inputting the hidden vector into a decoder in the abstract generation model for decoding to obtain a target abstract. The important keywords in the to-be-processed text are extracted preferentially, so that the model can identify key information of the to-be-processed text more effectively, and the accuracy of the generated target abstract is ensured.

The invention discloses an attribute-level sentiment analysis method based on a hierarchical attention mechanism and a gate mechanism, and aims to make full use of the relationship between contexts and attribute words to make the contexts and the attribute words fully associated. Highlighting the importance of different words in the context and the attribute words so as to improve the attribute-level sentiment analysis precision; the method enriches information in context and attribute word representation, and adopts the scheme that comment corpora are preprocessed, and a context word embedding matrix and an attribute word embedding matrix are obtained through GloVe word vector indexing; inputting the context and the attribute words into the GRU to obtain a context hiding state and an attribute word hiding state; obtaining a context vector representation 1, an attribute word vector representation and a context vector representation 2 through self-attention; splicing to obtain an overall vector representation, obtaining distribution of sentiment polarities, corresponding to contexts, of attribute words through a classifier, analyzing differences, and adjusting parameters in the model; the method belongs to the field of sentiment analysis in natural semantic processing.

The invention discloses a network traffic scheduling method based on deep reinforcement learning, and aims to provide intelligent routing and explore, utilize and learn an optimal path when a routing decision is made by utilizing network state information. According to the method, all link state indexes in the whole network are calculated by collecting network information, a reward function related to bandwidth, time delay and packet loss rate is designed, path state information is explored and learned by adopting a double deep Q learning network DDQN, and an optimal path between each pair of source and destination nodes is obtained. According to the method, an optimal route is installed on a switch of a data layer in advance by utilizing a network global view provided by an SDN and interaction between a DRL intelligent agent and an environment; the invention also provides a traffic scheduling system based on the method, which adopts a four-layer SDN structure comprising a knowledge layer, a management layer, a control layer and a data layer, and realizes dynamic scheduling of network traffic along with an actual environment.

The invention relates to the field of natural language technology processing, in particular to a dispute focus discovery method and device based on dispute focus entities and a terminal, and the method comprises the steps of obtaining document data, and carrying out the preprocessing of the document data to obtain an entity set; deleting redundant entities from the entity set to obtain a candidateentity set; splicing each candidate entity in the candidate entity set with the title and the document content to serve as an input feature; inputting the input features into a BERT model for training, and outputting a dispute focus entity after the training is completed; and performing dispute focus judgment according to the dispute focus entity. According to the invention, the candidate entity+ '-' + title + '-' + document content is used as an input feature, so that the attention of the candidate entity is higher; the binary classification task simplifies task steps of conventional entityidentification, each entity and a document can form a sample, the number of training samples is increased, meanwhile, the precision of dispute focus entities is greatly improved, and the effect is better.

The invention discloses an intelligent electrician practical operation examination system comprising a teacher computer, a tablet computer, a cabinet and multiple electrician practical operation examination system devices. That cabinet is internally provided with a power supply and a switch. The teacher computer communicates with each electrician practical operation examination system device through the Ethernet and the switch to send the instruction data to the electrician practical operation examination device and receive the result data of the electrician practical operation examination system. The electrician practical operation assessment system devices are mainly divided into a lighting circuit integrated practical training device and a drag circuit integrated practical training device. The advantages are that the examination computer performs automatic scoring according to the received signal without notary personnel for onsite detection so that the examination is more comprehensive, more efficient and fairer. After the examination, the examination computer automatically collects the data information of the operation safety behavior of students or examinees in the examination process to be uploaded to the server so as to establish a unique safety behavior file for later access and application of enterprises and schools.

The embodiment of the invention discloses a card information display method and device, computer equipment and a storage medium. The card information display method comprises the steps of receiving atrigger event, and collecting scene data according to the trigger event; determining a card display rule for card display in the current scene according to the trigger event and the scene data; acquiring card business data and/or card display resource data related to card display; and integrating the card business data and/or the card display resource data according to the card display rule, and generating card information in the current scene for display. The scene display rule is associated with a card corresponding to the scene; different data can be integrated and displayed only by switching different cards, the data can be conveniently displayed in a more centralized and more targeted mode, the displayed data are screened and integrated according to a display scene, the importance ofthe displayed data can be highlighted, the displayed data can be easily read by a user, and the information prompting efficiency and accuracy are improved.

A light stick for use in a traditional dance comprising an elongated impact resistant body having an striking surface and at least one light source imbedded in the body. An electronic circuit for controlling the light source and a power source for powering the electronic circuit and the light source may also be incorporated in the body. The light source is placed set back the said striking surface. The electronic circuit and the power source are contained within the body. A method of providing a special effect to a traditional dance comprises the step of illuminating a surface of the impact-resistant stick with imbedded lights is also described.

An attention model method for video surveillance pedestrian search based on natural language description is provided. The attention weights formed by fused feature vectors are used to weight the imagefeature vectors to obtain the degree of association between text and image. At the same time, the text is weighted by the attention weights formed by the text feature vectors, and the importance of the text is obtained. Attention evaluation score is obtained by multiplying the degree of connection between text and image by the importance of the text. The invention uses three different attention weighting modes, more effectively highlights the importance of each word text vector and improves the representativeness and effectiveness of the feature vector, so that the feature vector can be usedmore fully and effectively. The invention can make full use of the effective features in eigenvectors to improve the training and testing effect. The invention can solve some related problems such asartificial intelligence, machine learning and video surveillance pedestrian search based on natural language description.

The invention discloses a method and device for incoming call reminding of a smart phone, wherein the method and device achieve incoming call reminding through the smart phone with an OLED screen. The method comprises the following steps that S0, phone numbers stored in the smart phone are grouped to form at least one phone number set, and the corresponding relation is built between each phone number set and one area on the OLED screen; S1, when the smart phone is in a standby state with the OLED screen closed, and if incoming call exists, the phone number of the call is obtained; S2, whether the incoming phone number is a phone number in the phone number sets or not is detected; S3, pixel points in the area, having the corresponding relation with the phone number set to which the incoming phone number belongs, on the OLED screen are lightened. The method and device for incoming call reminding of the smart phone have the advantages that not only can system resources of the phone be utilized reasonably when the phone gets an incoming call, but also a user can intuitively know the identity of a caller according to the display areas and display modes of the OLED screen.

The invention relates to an image resolution determination method and device. The method comprises the following steps: obtaining a first image, the resolution of which is to be determined; determining a target area in the first image; and determining the resolution of the first image according to the resolution of the target area. Therefore, the method is capable of deciding the whole resolution of the first image through the resolution of the target area, so that the mistaken judgement of judging the first image as blurred picture when only the target area in the first image is clear is avoided, and then the speed and correctness of determining the resolution of the first image are improved.

The invention discloses a generalization safety evaluation method for a deep learning image classification model, and belongs to the technical field of machine learning. At present, the important problem to be solved in the deep learning related research is to improve the robustness of the model while solving the security threat problem with generalization characteristics faced by the deep learning image classification model, the invention utilizes the generalization security evaluation method oriented to the deep learning image classification model, by testing indexes such as the active defense capability, the adversarial sample detection capability and the passive defense capability of the deep learning image classification model for adversarial samples, the security of the deep learningimage classification model is comprehensively evaluated, security holes existing in the model are explored in the evaluation process, and meanwhile, due to the generalization characteristic of the method, the method can be suitable for most deep learning image classification models, and the method has important theoretical and practical significance for improving the safety of the deep learning field.

A vehicle-mounted sensor fused traffic target detection system of the present invention comprises sensors and an intelligent driving controller, the sensors comprise an RGB camera, a night vision camera, a millimeter wave radar and a laser radar, and the intelligent driving controller comprises a fusion algorithm module for outputting 3D target information, a motion planning module and a motion control module. The sensor inputs original data into the intelligent driving controller through a data line, and through a fusion algorithm of the intelligent driving controller, the control chassis is controlled through a CAN signal, and an automatic driving task is completed. Compared with a camera in the prior art, the robustness and the stability of the environment sensing module of the intelligent driving system can be improved.

The invention discloses application of microRNA-181s (miR-181s for short) in preparing a drug for treating acute enteritis. Specially, the effect of miR-181d is obvious. A fact that the small ribonucleic acid drug of miR-181d, which is modified with cholesterol, is concentratedly distributed in a small intestine is discovered through intravenous administration. The mechanism of action lies in that the small ribonucleic acid drug combines with complementary sequences in the TNF[alpha]3'-UTR area of a tumor necrosis factor to promote the degradation of TNF-[alpha]mRNA, so as to inhibit an acute enteritis reaction induced by TNBS. The small nucleic acid drug is obvious in effect, can be well concentrated in small intestine tissue, and enters cells to take effect. The small ribonucleic acid drug can also act on 3'-UTR of various inflammatory factors containing miR181d binding sequences to inhibit the expression of the inflammatory factors. Therefore, the small nucleic acid drug has high application value and good prospect for treating inflammatory bowel diseases.

The invention relates to the technical field of automatic driving vehicles, in particular to a three-dimensional reconstruction method for an intelligent vehicle based on multi-sensor fusion. The method comprises the following steps that camera image data and laser point cloud data are obtained through a camera and radar laser which are installed on a vehicle, then calibration of the camera, calibration of a radar and joint calibration of the camera and the radar are carried out, three-dimensional reconstruction is achieved according to the laser point cloud and the image data, data fusion of the radar laser and the camera is carried out, and finally, noise reduction and redundant information filtering after data fusion are carried out to obtain final three-dimensional reconstruction. Compared with traditional three-dimensional reconstruction, better three-dimensional reconstruction of the surrounding environment of the vehicle can be realized in a special environment, and a three-dimensional reconstruction effect with relatively high applicability and reasonability is realized, so that the landing process of the automatic driving vehicle is promoted, and the safety and reliability of the automatic driving vehicle are improved.