66results about How to "Improve method performance" patented technology

The invention relates to a deep attention mechanism-based image description generation method. The method comprises a deep long-short-term memory network model building step of adding an attention mechanism function between units of a long-short-term memory network model and training a long-short-term memory network added with the attention mechanism function by utilizing features and descriptioninformation of training images extracted by a convolutional neural network to obtain a deep long-short-term memory network model, and an image description generation step of generating descriptions corresponding to the images through a convolutional neural network model and the deep long-short-term memory network model in sequence according to the images to be subjected to description generation.Compared with the prior art, the method has the advantages of high information extraction effectiveness, high deep expression capability, accurate description and the like.

A process and system for dynamically compiling a partially interpreted method is provided. A set of bytecodes for a method is interpreted within a virtual machine. During the interpretation of the method, it is determined, according to the satisfaction of predetermined criteria, that the method contains an execution hot spot and should be just-in-time compiled (JITed) in order to increase the processing speed of the method. The interpretation of the method is halted with a halted execution state and at a halted execution location. Another method is constructed using information from the partially interpreted method and its execution state. The newly constructed method is just-in-time compiled and invoked in such a manner that the newly constructed method recreates the execution state of the partially interpreted method. Once the newly constructed method recreates the execution state of the partially interpreted method, the execution flow follows the bytecode sequence of the partially interpreted method.

A computer-implemented method, a system and computer programs products for assessing the credit worthiness of a user,the method comprising: collecting, by a data collector, information about communications conducted by users' of a communication network, the collected information at least comprising Call Detail Records including calls; analyzing, by a computing system, during a specific time frame, the collected information regarding a particular user including communications' started by the particular user and / or communications' received by the user, and determining, data variables from the analyzed information, the data variables including at least communication patterns; and computing, by the computing system, both a default risk score and a fraud risk score of the user by using the determined data variables.

The present invention generally relates to the detection of analytes, particularly biomolecules in samples. The invention also relates to compositions, methods, and kits for detecting the presence of analytes, typically in multiplex detection formats. The invention also relates to methods for determining the presence of at least one analyte in a sample, the methods employing employ single molecule detection techniques to individually detect at least one molecular complex or at least part of a molecular complex.

A method and an apparatus for identifying sub-graphs similar to a sample sub-graph from a network, in particular for identifying similar community structures from a network, are disclosed. Specifically, when a sample sub-graph is a community structure, it usually contains a plurality of nodes densely associated with each other. The method proposes first screening the network via key nodes of community structures so as to find sub-graph candidates that are potentially similar to the sample sub-graph, then strictly comparing the sub-graph candidates with the sample sub-graph to find similar sub-graphs of the sample sub-graph.

A decision feedback structure for recovering a bit stream out of received signals is disclosed, wherein the sampling instant may be tuned in dependence of the sequence or pattern of the preceding bits so as to follow the bit sequence dependent instant of the maximum eye opening. The decision-feedback equalizer structure includes decision means for making a bit value decision at a sampling instant, a feedback path to feed back bit values to said decision means and means for adapting the sampling instant for a bit value decision made by said decision means with respect to the sampling phase depending on the bit values of preceding bits, in particular depending on the bit value of the previous bit.

A technique for estimating channel data probability in a multi-user or multiple-input multiple-output communication system is disclosed. The technique uses parallel Markov Chain Monte Carlo simulation to select a plurality of hypothetical channel data patterns. Channel data bit probabilities are obtained by summing conditional bit probabilities, where the conditional bit probabilities are conditioned on an observation of the multi-channel signal and the hypothetical channel data patterns.

The device calculates a first frequency-dependent useful signal level estimator for the frame. The transfer function of a first noise-reducing filter is determined on the basis of the first useful signal level estimator and of a frequency-dependent noise level estimator. A second frequency-dependent useful signal level estimator for the frame is then calculated by combining the spectrum of the input signal and the transfer function of the first noise-reducing filter. The transfer function of a second noise-reducing filter is determined on the basis of the second useful signal level estimator and of the noise level estimator. The latter transfer function is used in a frame filtering operation to produce a signal with reduced noise.

A method and an apparatus for identifying sub-graphs similar to a sample sub-graph from a network, in particular for identifying similar community structures from a network, are disclosed. Specifically, when a sample sub-graph is a community structure, it usually contains a plurality of nodes densely associated with each other. The method proposes first screening the network via key nodes of community structures so as to find sub-graph candidates that are potentially similar to the sample sub-graph, then strictly comparing the sub-graph candidates with the sample sub-graph to find similar sub-graphs of the sample sub-graph.

The invention discloses a residual network and attention mechanism-based drug relationship extraction method. The method comprises the following steps of: S1, carrying out vector representation on words in a drug entity relationship data set; S2, carrying out time-series modeling on a drug relationship statement by utilizing a two-layer bidirectional long-short term memory model neural network; S3, importing residual connection into the constructed two-layer bidirectional long-short term memory model neural network; S4, decomposing a deep semantic meaning automatically obtained by the two-layer bidirectional long-short term memory model neural network into a memory space and an attention space, fusing memory information and attention information, and inputting the fused information into aSoftmax classifier to extract a drug relationship. According to the drug relationship extraction method disclosed by the invention, dependency relationships between long-distance words are effectivelysolved, gradient diffusion is overcome, model overfitting is prevented, the model robustness is good and the classification effect is good.

Disclosed is a method of detecting at least one oscillatory fault in at least one positional slaving chain for at least one airfoil of an aircraft. The method involves estimating a reference position of the airfoil in the absence of a fault, and calculating a residual value based on the difference between the estimated reference position and the actual position measured by at least one sensor. The calculated residual value is compared with at least one predetermined threshold value to determine the number of successive and alternating overshoots of the predetermined threshold value by the residual value, and oscillatory fault is determined based on the determined number.

A method for modified dialing behavior includes capturing a dialing stream in a virtual input buffer, comparing the dialing stream of the calling party to a history of dialing streams stored in a database, analyzing the history of dialing streams to determine a desired dialing behavior of the calling party, and passing a modified dialing stream to a mobile operating system based on the determined desired dialing behavior of a calling party.

The invention relates to a method for a medical image analysis, comprising the steps: Performing a first CT scan of a region of interest of a subject resulting in a first image with a first resolution; Applying a medical image processing method to the first image resulting in first values representing a first analysis of the region of interest of the subject; Determining a range of interest of the subject based on the first values; Performing a second CT scan of the range of interest of the subject resulting in a second image with a second resolution, wherein the second resolution is higher than the first resolution; and Applying the medical imaging processing method to the second image resulting in second values representing a second analysis of the range of interest of the subject. Further, the invention relates to a system for medical image analysis.

A device for polishing centrifugal impellers for a turbomachine including a vat configured to be filled with a polishing agent, and an impeller support configured to make the impeller rotate around its axis and move it along its axis such that all of points of the impeller have a helical movement whereof the pitch is close to that of the helix from which the general shape of the airflow channels of the impeller comes, delimited by the blades of the impeller.

An electronic method for non-target multi-index rapid detection of pesticide residues in edible agricultural products is disclosed. The method includes 1) building a primary accurate-mass mass spectrum database and a secondary fragment ion mass spectrum database for 1200 or above pesticides commonly used in the world, applying the two databases for pesticide compound screening and identification,and establishing a unique electronic identity for each pesticide so that traditional identification methods adopting pesticide material standards as references are replaced by electronic standards; 2)developing intelligent screening software for automatic matching between a high resolution mass spectrum and a pesticide residue mass spectrum and qualitative identification, implanting the softwareprogram into an instrument to achieve high-speed, high-throughput, high-precision, high-reliability, highly informationized and highly automatic comparison of a test sample with pesticide mass spectrum databases; and 3) preparing samples by one time, and rapidly detecting 1200 pesticides simultaneously by adopting two detection techniques. Non-target multi-index rapid detection of pesticide residues in edible agricultural products is achieved, and the detection efficiency is greatly increased.

A method for controlling power flow within a DC power flow within a DC power transmission network includes two or more interconnected converter stations. The method includes the steps of: establishing a common feedback signal (Ud,common, Ũd,common) to the converter stations, the common feedback signal (Ud,common, Ũd,common) being based on an overall voltage level in the DC transmission, power network; providing, in the converter stations, a control signal (S1, S2, S3, S4) based on the common feedback signal (Ud,common) and a drooped error signal (Pe, droop, Ue, droop); and controlling the power flow within the DC power transmission network towards set operating points (Udref, PCCref) by using the control signal (S1, S2, S3, S4). A control device and computer program product is also described.

The present invention relates to a ground milling machine, particularly a road milling machine, comprising a machine frame supported by travelling devices, a drive engine and a milling device arranged in a milling drum box, the milling drum box comprising a stripping plate which is height-adjustable by means of at least a first and a second actuator. Further, the present invention relates to a method for adjusting a stripping plate of a milling drum box of a ground milling machine, particularly a road milling machine. For the purpose of a reliable adjustment of the stripping plate, the present invention proposes using a monitoring device for detecting the adjustment of a first actuator and a control device for controlling the adjustment of a second actuator depending on the adjustment of the first actuator.

The invention discloses a temperature and strain correlation-based bridge damage identification method. The method comprises the steps of 1) performing abnormal value removal on bridge health monitoring data and extracting a strain component influenced by temperature; 2) fitting temperature and strain component data by using a linear regression model, and extracting a corresponding regression coefficient; 3) performing normal check and normal conversion operation on the regression coefficient, and then performing statistics on a mean value mu and a standard deviation sigma; and 4) preprocessing latest monitoring data, performing fitting by using the linear regression model to obtain a regression coefficient, and according to a 3 sigma criterion, judging whether a bridge is damaged or not.The method is simple and efficient in principle, and does not involve excessively complex algorithms; a large amount of data (the required data directly comes from an existing bridge health monitoringsystem, additional equipment such as a sensor and the like does not need to be added, and the bridge operation cost is reduced) collected by the monitoring system can be effectively utilized; and thecontinuously accumulated monitoring data can further improve the effect of the method.

Provided is a computer-implemented method for inter-sententially determining a semantic relationship between a first entity and a second entity in a natural language document, comprising at least the steps of:generating a first dependency parse tree, DPT, for a first origin sentence of the document which comprises the first entity, wherein each DPT comprises at least a root node;generating a second DPT for a second origin sentence of the document which mentions the second entity;linking the root nodes of the first DPT and the second DPT so as to create a chain of words, COW;determining for each word in the COW a subtree;generating for each word in the COW a subtree embedding vector cw which is based at least on word embedding vectors xw of the words of the subtree;generating a representation vector pw for each word in the COW; andclassifying, using a recurrent neural network, the semantic relationship between the first entity and the second entity, based on the input representation vectors pw.

A method for managing a closed incisional wound that includes at least one elastomeric drain inserted partially into the wound, an open cell foam sponge in flow communication with the drains and wound, and to which negative pressure is applied by a vacuum source. The negative pressure is transmitted into the otherwise closed wound in order to improve drainage and expedite wound healing.

The invention discloses a geometric shadow map method for triangle reconstruction, and the method comprises the steps: (1) drawing a scene in a light source space, and generating and storing a geometric shadow map; (2) carrying out the compression of the obtained geometric shadow map based on a GPU, and obtaining a compressed geometric shadow map; (3) drawing a scene in a vision field, packaging pixels, carrying out triangle consistency check, and using a stored triangle to achieve the geometric reconstruction of the depth value of a pixel shading object; (4) comparing a pixel depth value in the light source space with the depth value of the shading object, obtaining a shadow calculation result, and completing the reconstruction of the geometric shadow map. Aiming at problems of deformation and depth deviation in a shadow map method, the method employs discrete scene raw triangular patches to represent and record the information of the shading object, and the invention provides the geometric shadow map method. While generating a high-quality shadow, the method has certain advantages over other algorithms in geometric data storage and expenditure and performance, and can meet the demands of real-time application better.

The invention discloses a video frame rate promoting method, which comprises the following steps: reading in a video; setting a counter i=1; calculating gradient values of all pixel points in a ith frame as the current frame and the next frame of the current frame of the video; taking gradient values of the pixel points in the ith frame and the i+1th frame as pixel values; reconstructing a pixel point set of the ith frame and the i+1th frame to form the reconstructed ith frame and i+1 frame; carrying out bidirectional layering motion estimation on the reconstructed ith frame and i+1 frame based on forecast, so as to obtain a forward motion vector field and a backward motion vector field; calculating the motion vector of each pixel in the ith frame and the i+1 frame by adopting bilinear interpolation according to the forward motion vector field and the backward motion vector field; and carrying out occlusion judgment on the pixels in the ith frame. Compared with the existing video frame rate promoting method, the video frame rate promoting method disclosed by the invention has the advantages that the influence of the quality of the video on motion estimation accuracy is reduced, the problem brought by over-large scale partitioning is solved, and the result of the bilinear interpolation is more accurate.

The invention discloses a chaotic encryption method based on Logistic mapping, which is applied to video encryption and decryption transmission. A server and a client are included, the server is usedas a sending end, the client is used as a collecting end, and the method comprises the following steps: 1, collecting a video and capturing a video frame by the sending end; 2, setting a video frame acquisition frame rate, and encoding the acquired video frames; 3, encrypting the encoded video by using a chaotic encryption algorithm of Logistic mapping; 4, establishing a real-time transmission network transmission protocol RTP session, and sending encrypted data; 5, creating a thread by the collection end, and receiving a data packet by using a network transmission protocol RTP session for creating real-time transmission; step 6, decrypting the received video by using a Logistic mapping chaotic encryption algorithm; 7, decoding the received video; and 8, displaying the video frame by usinga picture display control.

Aldehyde- and ketone-functionalized proteins are promising new substrates for the development of chemically modified biotherapeutics and protein-based materials. Their reactive carbonyl groups are typically conjugated with a-effect nucleophiles, such as substituted hydrazines and alkoxyamines, to generate hydrazones and oximes, respectively. However, the resulting C═N linkages are susceptible to hydrolysis under physiologically relevant conditions, which limits their utility in biological systems. Here we introduce a Pictet-Spengler ligation that is based on the classic Pictet-Spengler reaction of aldehydes and tryptamine nucleophiles. The ligation exploits the bioorthogonal reaction of aldehydes and alkoxyamines to form an intermediate oxyiminium ion; this intermediate undergoes intramolecular C—C bond formation with an indole nucleophile to form an oxacarboline product that is hydrolytically stable. The reaction was utilized for site-specific chemical modification of glyoxal- and formylglycine-functionalized proteins, including an aldehyde-tagged variant of the therapeutic monoclonal antibody Herceptin. In conjunction with techniques for site-specific introduction of aldehydes into proteins, the Pictet-Spengler ligation offers a new means to generate stable bioconjugates for medical and materials applications.

The invention provides a session management method. The session management method comprises the following steps of: taking nginx as a reverse proxy, and developing a nginx plug-in to provide a cluster front end of an application portal; using Ivs (Linux Virtual Server) to build a high-availability cluster based on reverse proxy software of the nginx; using cluster redis as a uniform session for storage; developing a session management module in the reverse proxy; and stripping session data and an application server. The session management method has the advantages of high execution efficiency, high availability and low cost; through expired processing of the session and the stripping from the application server, the problem of desynchrony of the sessions among application server clusters can be effectively avoided, the application servers are independent to one another, the session synchrony is not needed, and better elasticity and performance can be provided.

The invention discloses a multipath transmission routing method based on an immune particle swarm. The method comprises the following steps: dynamically establishing a plurality of disjoint transmission paths between a source node and a destination node in a network area where mobile sensing nodes are deployed; using an improved immune particle swarm method based on adaptive search to group sub-populations, dynamically adjusting the scale of each group of sub-populations; in combination with the concentration adjustment mechanism, adaptively adjusting the search range according to the maximumconcentration value of the particles; vaccinating the suboptimal population; using a particle maximum concentration value to adjust a search range of vaccination. Population degradation is avoided, convergence precision and global search capability of the method are improved, the method has good convergence precision and global search performance, method performance is improved, and influence on transmission stability and reliability of wireless sensor network routing caused by fault nodes is avoided.

The invention relates to a puerarin injection and a quality control method thereof. 1Ml of the puerarin injection comprises 10-100mg of puerarin, 400-600mg of propylene glycol, and injection water added to an equilibrium volume. The method using HPLC to detect relevant substances in the puerarin injection has an excellent methodological performance.

The present invention relates to a method and system for controlling a number of writing cycles supported by a cell or portion (11) of a non volatile memory (4) of an IC Card (10), including the steps of counting write accesses to the memory portion (11) and storing a first counter (21) of the write accesses in another portion (21) of said non volatile memory (4). The method comprises coupling the first counter (21) to a second counter or value (31) associated to a RAM (4) (Random Access Memory) of the IC Card (10), wherein the second counter or value (31) is updated each time the write accesses occur on said cell or portion (11) to be controlled and the first counter (21) is written in the another portion of non volatile memory only when the second counter or value (21) corresponds to a predetermined value.