Patents
Literature
Hiro is an intelligent assistant for R&D personnel, combined with Patent DNA, to facilitate innovative research.
Hiro

36results about How to "Effective excavation" patented technology

Face detection method, and training method and device of face detection model

The invention discloses a face detection method for protecting user privacy and improving information security. The method can be applied to the field of artificial intelligence, and comprises the following steps: acquiring a face image sequence; obtaining a noise image sequence according to the face image sequence; obtaining a face feature map sequence of the face image sequence through a first feature extraction network of a face detection model; obtaining a noise feature map sequence of the noise image sequence through a second feature extraction network of the face detection model; based on the face feature map sequence and the noise feature map sequence, obtaining a classification probability value corresponding to the face image through a full connection layer of the face detection model; and determining a detection result according to the classification probability value. The invention further provides a training method and device of the face detection model. By paying attentionto semantic information of the human face image sequence and noise information of the noise image sequence at the same time, artifacts generated by fake human faces can be effectively mined, and theaccuracy of human face detection is improved.
Owner:TENCENT TECH (SHENZHEN) CO LTD

Deep learning prediction method based on combined framework of plurality of convolution neural networks

The invention belongs to the technical field of data identification and data representation and discloses a deep learning prediction method based on the combined framework of a plurality of convolution neural networks. According to the invention, the method comprises the steps of carrying out the dimension transformation of data; preparing a training data set, a verification data set and a test data set; inputting the data subjected to dimension transformation correspondingly into different convolution neural networks according to periodic state numbers, rearranging the data processed by the convolution neural networks according to the time sequence and then inputting the data into a deep full-connection neural network to obtain a final result; performing early termination training by using the verification data set to obtain a model; predicting the test set to obtain a prediction result. According to the invention, the data of different periodic states are respectively processed through the plurality of convolution neural networks, so that the data rule information is mined in a targeted mode. The number of network layers is reduced. The rearrangement sequence of data after the treatment of convolution neural networks is carried out and then the data are input into the deep full-connection neural network. The data dimension is greatly reduced. The over-fitting is effectively relieved, and the prediction accuracy is improved.
Owner:XIDIAN UNIV

News recommendation method based on knowledge perception

The invention discloses a news recommendation method based on knowledge perception, and belongs to the technical field of artificial intelligence and network information big data. The method comprisesthe following steps: firstly, identifying entities from a news text, matching the entities with entities in a knowledge graph, and obtaining vector representation of the entities and relationships byutilizing a knowledge representation learning model; secondly, fusing the word vectors and the entity vectors of the news through a convolutional neural network, and learning comprehensive features of the news from the word vectors and the entity vectors; for each piece of candidate news, dynamically constructing user features by using an attention network based on the correlation between the user browsing history and the candidate news, meanwhile, considering paths between entities in clicked news and entities in candidate news, and generating path features from the paths through a long-termand short-term memory network; finally, taking the candidate news features, the user features and the path features as input, and calculating the probability that the user clicks the candidate news by using a multi-layer perception mechanism to achieve a recommendation purpose. Knowledge association between news can be effectively mined, and the method is remarkably superior to a traditional method.
Owner:BEIJING INSTITUTE OF TECHNOLOGYGY

Concave cutter of slurry balance tube push bench

InactiveCN103161472AAdapt to excavation requirementsImprove stabilityTunnelsSlurrySpoke
The invention relates to a concave cutter of a slurry balance tube push bench. The concave cutter comprises a cutter body, main cutter blades and first-moving blades. The cutter body is composed of spokes, a side plate, a rib plate, an outer edge plate, a rear cover plate, an abrasion resistance alloy strip and a support beam in a welding mode. The spokes are in a concave shape, the spokes are integrally connected through the side plate, the rib plate and the outer edge plate, and each spoke is provided with the main cutter blade and the first-moving blade. An inner cutter body is arranged inside the cutter body, and an outer cutter body is arranged outside the cutter body. A digging cutting track of the inner cutter body is provided with two main cutter blades and two first-moving blades in an assembling mode. A digging cutting track of the outer cutter body is provided with three main cutter blades and three first-moving blades in an assembling mode, and an outermost digging cutting track is provided with twelve scrapers in an assembling mode. The concave cutter of the slurry balance tube push bench can be suitable for the digging demands of complex soil layers and has the advantages of being high in opening rate, capable of supporting and protecting to a digging face, good in abrasion resistance performance and capable of crushing rock soil at second time and the like.
Owner:SHANGHAI FOUND ENG GRP

Round cutter box capable of having disc cutter replaced bidirectionally

The invention relates to the technical field of tunnel boring machine engineering machinery, in particular to a round cutter box capable of having a disc cutter replaced bidirectionally. The round cutter box capable of having the disc cutter replaced bidirectionally comprises a ringlike cutter box body and two cutter shaft fastening units, wherein the ringlike cutter box body is fixed on a cutter disc and is in a cylindrical shape, the cutter shaft fastening units are arranged at the inner side of the cutter box body and are in central symmetry, and each cutter shaft fastening unit comprises a load-bearing guide bar, a limiting guide bar, a wedge-shaped pressing block, a tensioning block, a tension bolt and nuts, wherein each load-bearing guide bar is fixed to the inner wall of the cutter box body, each limiting guide bar is fixed to the inner wall of the cutter box body and arranged in a mode that the limiting guide bar corresponds to the corresponding load-bearing guide bar and a certain included angel is formed between the limiting guide bar and the corresponding load-bearing guide bar, and each wedge-shaped pressing block, the corresponding tensioning block, the corresponding tension bolt and the corresponding nuts are arranged between the corresponding load-bearing guide bar and the corresponding limiting guide bar in a matched mode. By means of the round cutter box, the disc cutter can be replaced from either the back side or the front side of the cutter disc, the cutter is replaced from the back side in the construction process of a tunnel boring machine, the cutter is replaced from the front side in the production, assembling and debugging process of the tunnel boring machine, and the cutter box body of the round cutter box is of a round structure, so that the cutter box body on the cutter disc is made to be capable of saving the space and being subjected to even stress.
Owner:CHINA RAILWAY ENGINEERING EQUIPMENT GROUP CO LTD

Intelligent metal detection device

The invention discloses an intelligent metal detection device, and the device comprises a detection module, a control module, a Bluetooth communication module, an intelligent terminal, and a cloud server. The detection module emits an electromagnetic wave, and senses an electromagnetic signal transmitted by underground metal. The control module carries out the data processing of the electromagnetic signal, and processes the data into data information intelligently. The Bluetooth communication module communicates with the intelligent terminal. The intelligent terminal is provided with a metal detection application program, and the application program generates a treasure detection signal overlapping track map through combining with the signal transmitted by the Bluetooth communication module, the GPS track function of the intelligent terminal and map software. The application program also can receive the metal detection device hardware updating information pushed by a manufacturer network, and achieves the direct updating of the hardware. According to the invention, the device provided by the invention can achieve the network sharing of the detection information of metal, facilitates the interactive exchange of metal detection operation and effective exploration of metal articles, can receive the hardware updating information pushed by the network for direct updating, and saves the time and logistics cost in returning to a factory for hardware updating.
Owner:SHENZHEN TUOMAI ELECTRONICS CO LTD

Dissolved oxygen concentration soft-measuring method based on group decision case reasoning

The invention provides a dissolved oxygen concentration soft-measuring method based on group decision case reasoning, realizes the precise estimate of dissolved oxygen concentration, and aims to solve the problems that an analytical instrument for dissolved oxygen concentration in the sewage disposal process lags in measurement and difficult to built a precise mathematic model for estimate the dissolved oxygen concentration. According to the method, a case reasoning soft-measuring model can be built through the steps of determining characteristic variable, building soft-measuring case library, distributing weight, obtaining target cases, normalizing numerical value, retrieving cases, reusing group decision cases, storing cases, and the like, and an algorithm can be implemented, so that the measurement error can be reduced; the problem of measuring lag of the analytical instrument can be solved, and the complex process for building a precise mathematic model can be avoided; the experimental result shows that the method can be used for quickly and precisely estimating the dissolved oxygen concentration, has the self-learning ability, reduces the purchasing and maintenance cost of the analytical instrument, and facilitates the real-time monitoring of the dissolved oxygen concentration.
Owner:BEIJING UNIV OF TECH

Prefabricated retaining wall and pile hole excavation method

PendingCN110438988ATo meet the needs of hole formationThe settlement process is easyExcavationsBulkheads/pilesLarge sizeInternal stress
The invention discloses a prefabricated retaining wall and a pile hole excavation method. The excavation method comprises the steps that first, guide-hole construction is carried out by using a machinery in an excavation area of a rectangular hole, after the internal opening is carried out, a soil body at the edges and corners of the rectangular hole is disturbed, internal stress is released, thedifficulty of subsequent excavation is greatly reduced, moreover, with the continuous increase of gravity of the prefabricated retaining wall, the settlement process of the prefabricated retaining wall is easier, and construction efficiency is conducive to be improved; and according to the excavation method, excavation of the rectangular hole can be carried out effectively, construction safety isimproved, construction difficulty is reduced, the requirement for forming a large-size rectangular hole is met, at the same time, prefabricated retaining wall structure is effectively installed to avoid the risk of hole collapse, hole digging and the retaining wall can be completed simultaneously, the construction period is conducive to be shortened, energy saving and environment production are achieved, the quality hidden danger of the retaining wall can be avoided effectively, and the excavation method is especially suitable for construction of an anti-slide pile in cold and high altitude areas such as the sichuan-tibet line.
Owner:CHINA RAILWAY ERYUAN ENG GRP CO LTD

Collaborative relation graph-based recommendation method and related device

The invention discloses a recommendation method based on a collaborative relation graph and a related device. The recommendation method comprises the steps that the collaborative relation graph is constructed, interaction nodes with entity nodes are obtained based on the collaborative relation graph, and implicit relation embedding representation of the entity nodes is determined based on the interaction nodes. Obtaining triples taking the entity nodes as head entities on the basis of the collaborative relation graph, and determining explicit relation embedding representation of the entity nodes on the basis of all the obtained triples; obtaining high-order domain information of the entity node based on the implicit relation embedding representation and the explicit relation embedding representation; and training a recommendation model based on the high-order domain information, and determining an interaction probability between the to-be-recommended user and each candidate item corresponding to the to-be-recommended user through the recommendation model. According to the method, the user-article bipartite graph and the knowledge graph are integrated, and meanwhile, the implicit relationship between the user and the article and the attribute relationship between the entities are utilized to carry out high-order information propagation, so that the high-order neighborhood information of the nodes is effectively explored, and the recommendation accuracy is improved.
Owner:SHENZHEN UNIV

Grooving device for aluminum profile plate machining

The invention discloses a grooving device for aluminum profile plate machining, and belongs to the field of plate machining. The grooving device for aluminum profile plate machining is disclosed and comprises a supporting frame, a supporting plate and a fixing base. The output end of a first driving motor is connected with a first rotating device, the end, away from the first driving motor, of the first rotating device is connected with a first sliding assembly, the output end of the first sliding assembly is connected to the top wall of the supporting plate, a second driving motor is connected to the outer wall of the second connecting plate, a second rotating device is connected to the output end of the second driving motor, an air cylinder, a first fixing plate and a sewage tank are connected to the bottom wall of the supporting plate, and a fixing device is connected to the output end of the air cylinder. The grooving device is simple in structure and convenient to operate, and is capable of effectively cutting through any position on the surface of an aluminum profile plate; and chippings generated after the aluminum profile plate material is cut through can be cleaned away by arranging a cleaning brush and a draught fan, and are thus prevented from polluting the environment, and the flexibility and practicability of the device are further improved.
Owner:福建中创新型材料科技有限责任公司

Mongolian-Chinese neural machine translation method based on multiple constraint terms

PendingCN114818743AAlleviate the problem of insufficient trainingSolve the problem of poor fluencyNatural language translationNeural architecturesAlgorithmWord list
A Mongolian-Chinese neural machine translation method based on multiple constraint terms comprises the steps that firstly, a model training process based on reinforcement learning is constructed for Mongolian-Chinese neural machine translation tasks, and then constraint conditions are further improved for a training optimization target on the basis of a reinforcement model; comprising the following steps: adding a semantic constraint module to relieve the problem of poor translation fluency caused by a single BLEU value evaluation system; performing parameter constraint on the training process to improve the efficiency of model training; and performing word list constraint on the corpus to reduce the number of unregistered words in the translation. According to the method, by adjusting the overall constraint mode, the problems that in a low-resource Mongolian-Chinese machine translation task, a model is poor in sequence structure analysis capacity and low in training efficiency are solved, and meanwhile the problem that variance is large due to intensive training is solved; according to the method, a mean value reward and pruning column search method is adopted to effectively relieve the negative influence caused by the enhanced training.
Owner:INNER MONGOLIA UNIV OF TECH

Multi-view semi-supervised classification method for rapid seed random walk

The invention relates to a multi-view semi-supervised classification method for rapid seed random walk, which comprises the following steps: firstly, calculating a similarity matrix and a transition probability matrix of each view of input multi-view data by adopting a Gaussian kernel function; secondly, establishing an initial distribution state of each view angle of the multi-view-angle data according to the category label of the multi-view-angle data for semi-supervised learning, and calculating an arrival probability matrix of a first transfer state of each view angle of the multi-view-angle data; and finally, iteratively calculating an arrival probability matrix of multiple transition states of each view angle of the multi-view-angle data, and performing weighted summation on the arrival probability matrixes of all the transition states of each view angle to obtain a reward matrix of each view angle so as to generate a category label of the multi-view-angle data for testing. According to the invention, various types of data such as images, texts and videos can be accurately and effectively classified only by using a small amount of supervision information, and the method has acertain practical value.
Owner:FUZHOU UNIV

Special excavating equipment applicable to underground excavation construction of underground engineering

The invention discloses special excavating equipment applicable to underground excavation construction of underground engineering. The special excavating equipment comprises a supporting mechanism, an excavating mechanism, a prefabricated part assembling mechanism and a transporting mechanism which is used for transporting a soil body and a prefabricated part. The special excavating equipment has the beneficial effects that (1) the complexity degree and equipment cost of the excavating equipment are greatly reduced as compared with those of the traditional shield machine; (2) the excavating equipment has a special-shaped excavating cross section, can be applicable to different construction environments and excavating cross sections, can effectively excavate a shallow soil coverage area (generally 2 to 3 m in depth) of an urban trunk road, and meanwhile, can precisely determine an excavating angle according to a rupture angle of a rock soil body; (3) the excavating equipment is developed relying on an innovative underground engineering construction method, and can realize the functions of top plate supporting, cross section excavating, prefabricated part assembling, muck and prefabricated part transporting, synchronous grouting and the like in a construction process.
Owner:JINAN RAILWAY TRANSPORT GRP CO LTD +1

Method for automatically testing performance of windows desktop software

The invention discloses a method for automatically testing the performance of windows desktop software, and belongs to the technical field of software testing. The method comprises the following steps: 1, acquiring a test case set, and setting an iterative threshold; 2, recording the probability of switching between any two test cases as Pij, and establishing an array for storing the probability of switching between all nodes; 3, initializing all nodes, and initializing the pheromone strength among the nodes to be 1 and the pheromone among the nodes to be 0; 4, randomly selecting a group of test case combinations, and updating pheromones according to the change of system resources after the execution of the test case combinations is completed; 5, if resource growth occurs, additionally increasing pheromones among the current nodes, and updating global pheromones; 6, when the pheromone is higher than the threshold value, entering a confirmation link, and repeatedly executing the path inthe confirmation link to update the pheromone according to the judgment result; and 7, terminating the test after the number of algorithm iterations exceeds a threshold, and outputting a test conclusion.
Owner:SICHUAN AEROSPACE SYST ENG INST

A news recommendation method based on knowledge perception

The invention discloses a news recommendation method based on knowledge perception, which belongs to the technical field of artificial intelligence and network information big data. Firstly, entities are identified from news texts, matched with entities in the knowledge graph, and vector representations of entities and relationships are obtained using the knowledge representation learning model. Then, a convolutional neural network is used to fuse word vectors and entity vectors of news, and learn comprehensive features of news from them. For each news candidate, based on the correlation between the user's browsing history and the news candidate, user features are dynamically constructed using an attention network. At the same time, considering the path between the entity in the clicked news and the entity in the candidate news, the long short-term memory network is used to generate path features from the path. Finally, with the candidate news features, user features and path features as input, the multi-layer perception mechanism is used to calculate the probability of the user clicking on the candidate news to achieve the purpose of recommendation. The invention can effectively excavate the knowledge association between news, which is significantly better than the traditional method.
Owner:BEIJING INSTITUTE OF TECHNOLOGYGY

A treatment method for poor foundation of concrete dam joints

The invention relates to a treatment method for poor concrete dam joint foundations. The treatment method for the poor concrete dam joint foundations does not affect operation of railways or highways and meets the requirements of water-power engineering junction layouts. The treatment method aims at achieving the foundation strength and the seepage-proofing purpose of a concrete dam joint under poor geological conditions, guaranteeing operation safety of the highways in the flood season, reducing slope excavation, reducing waste slag fields, facilitating water and soil conservation and saving engineering investment. According to the technical scheme, the concrete dam joint is arranged on one side of a highway and comprises the concrete dam joint foundation I adjacent to the highway and the adjacent concrete dam joint foundation II; after the concrete dam joint foundation I is excavated, steep slopes are formed, plain concrete 20 cm in thickness needs to be sprayed to the steep slopes for slope fixation, it is found that a foundation weak intercalated layer is arranged below the concrete dam joint foundation I during construction with the attitude of rock slowly inclines to the downstream, and the inclination angle ranges from 20 degrees to 35 degrees. The treatment method is suitable for hydraulic and hydroelectric engineering.
Owner:POWERCHINA HUADONG ENG COPORATION LTD

A face detection method, face detection model training method and device

This application discloses a face detection method for protecting user privacy and improving information security, which can be applied to the field of artificial intelligence. This application includes: obtaining a sequence of human face images; obtaining a sequence of noise images according to the sequence of human face images; The first feature extraction network of the face detection model obtains the face feature map sequence of the face image sequence; the noise feature map sequence of the noise image sequence is obtained through the second feature extraction network of the face detection model; based on the face feature map sequence and Noise feature map sequence, the classification probability value corresponding to the face image is obtained through the fully connected layer of the face detection model; the detection result is determined according to the classification probability value. The present application also provides a training method and device for a face detection model. The present application pays attention to the semantic information of the face image sequence itself and the noise information of the noise image sequence at the same time, so as to effectively discover artifacts produced by forged faces and improve the accuracy of face detection.
Owner:TENCENT TECH (SHENZHEN) CO LTD

Large-section soft ground tunnel full-section construction structure and construction method

The invention discloses a large-section soft ground tunnel full-section construction structure and a construction method. The structure comprises a tunnel excavation and anchorage system formed by multiple tunnel excavation and anchorage mechanisms for excavating and anchoring rock masses in front of a working face, wherein each tunnel excavation and anchorage mechanism comprises an anchor rod, ananchorage mechanism and a rotary excavation cutter. The method comprises the steps of (1) constructing a rear-end segment; (2) constructing a next tunnel segment; and (3) repeating the step (2) for one time or multiple times until the tunnel construction process is accomplished. The large-section soft ground tunnel full-section construction structure provided by the invention is reasonable in design, simple and convenient to construct and good in use effect; the construction quality is easy to control; and when the rock masses in front of the working face are effectively dynamically anchoredby the tunnel excavation and anchorage system formed by adopting the multiple tunnel excavation and anchorage mechanisms, a large-section soft ground tunnel can be mechanically full-section excavatedat the same time, so that the tunnel surrounding rock and tunnel structure is ensured to be stable, the construction risk is reduced, the construction progress can be effectively accelerated, and thetunnel construction quality is easy to control.
Owner:CHINA RAILWAY FIRST GRP CO LTD

Information generation method and cloud service platform based on intelligent online communication and big data

The embodiment of the present application provides an information generation method and cloud service platform based on intelligent online communication and big data. Considering that the online communication fault interaction behavior is characterized by merging the online communication fault data into the original service access big data and the associated The data recording area corresponding to the service access big data, therefore, the online communication failure data in the data recording area corresponding to these original service access big data will produce some specific data recording areas inconsistent with the original access data. Based on this, attention should be paid to the service access data. The characteristic information of the data collection itself and the characteristic information of the online communication failure data collection can effectively discover the optimized communication interaction objects corresponding to the service access big data collection, so as to facilitate subsequent communication optimization for the optimized communication interaction objects in a timely manner.
Owner:江苏臻云技术有限公司

Passenger flow time sequence clustering optimization flight regulation and control method and system

The invention relates to a passenger flow time sequence clustering optimization flight regulation and control method and system, and the method comprises the steps: S1, carrying out the grouping statistics according to flight historical orders, and obtaining flight data; S2, performing normalization processing on the flight data to obtain normalized flight data; S3, calculating KL divergences of the normalized flight data in pairs according to dates, and distributing the normalized flight data to a preset interval according to the KL divergences; and S4, clustering the normalized flight information of each interval to obtain the category of each piece of normalized flight information, and taking the category as reference information for optimizing flight regulation and control. According to the method provided by the invention, similar conditions of the aviation historical orders can be effectively mined, the similar conditions are provided for flight planning personnel as reference for optimizing flight regulation and control, the planning cost is saved, and the efficiency is improved.
Owner:北京人人云图信息技术有限公司

Treatment method for poor concrete dam joint foundations

The invention relates to a treatment method for poor concrete dam joint foundations. The treatment method for the poor concrete dam joint foundations does not affect operation of railways or highways and meets the requirements of water-power engineering junction layouts. The treatment method aims at achieving the foundation strength and the seepage-proofing purpose of a concrete dam joint under poor geological conditions, guaranteeing operation safety of the highways in the flood season, reducing slope excavation, reducing waste slag fields, facilitating water and soil conservation and saving engineering investment. According to the technical scheme, the concrete dam joint is arranged on one side of a highway and comprises the concrete dam joint foundation I adjacent to the highway and the adjacent concrete dam joint foundation II; after the concrete dam joint foundation I is excavated, steep slopes are formed, plain concrete 20 cm in thickness needs to be sprayed to the steep slopes for slope fixation, it is found that a foundation weak intercalated layer is arranged below the concrete dam joint foundation I during construction with the attitude of rock slowly inclines to the downstream, and the inclination angle ranges from 20 degrees to 35 degrees. The treatment method is suitable for hydraulic and hydroelectric engineering.
Owner:POWERCHINA HUADONG ENG COPORATION LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products