61results about How to "Avoid prediction" patented technology

The invention discloses a multi-form bus dynamic scheduling method based on bus station informatization. The method comprises steps that (1), connection and interaction between the bus station information and the information of a passenger handset terminal are carried out to acquire various data; (2), an initial bus scheduling model is established, and model solution is carried out to acquire optimal bus initial departure time intervals in multiple forms; (3), each coefficient is calculated, and a scheduling form and an operation section of each two adjacent buses are determined; (4), a real-time multi-form bus scheduling model is established, and the model is solved to acquire the optimal bus station quantity between each two adjacent buses and the optimal operation time among the bus stations; and (5), sum of optimal operation time intervals between each two adjacent buses is calculated according to time difference. The method is advantaged in that the bus station information technology and the information of the passenger handset terminal are fused, dynamic scheduling in different bus forms and with optimal bus operation intervals is realized, and bus operation efficiency is improved.

The invention discloses a dynamic common bus scheduling method based on bus station informatization. The method comprises: (1), various data are obtained based on connection and interaction between bus station information and passenger phone terminals; (2), a real-time bus scheduling model is established and is solved, thereby obtaining an optimal station number between each two adjacent buses and optimal operation time between the stations; and (3), optimal operation time intervals of adjacent vehicles are obtained by summation of the time differences. Therefore, fusion of the bus station informatization technology and the passenger phone terminal information is realized, thereby realizing dynamic scheduling of the bus form and the optimal bus-to-bus interval; the bus operation efficiency is improved; the organic virtuous circle of the public transit system of the city is realized, so that residents or passengers can select the optimal bus way conveniently for trip and thus the bus attraction can be enhanced finally.

The invention provides a face alignment method and device. The face alignment method and device are used for relieving the problems that during face alignment, the calculated amount is large, the requirement on hardware is high, and faces are hardly transplanted to mobile equipment. The method includes the steps that multiple face alignment models with different deflection angles are established;when the face alignment models with different deflection angles carry out prediction of T stages on a to-be-detected face image, the confidence coefficient of prediction shape, predicted on the current prediction stage by the face alignment models with different deflection angles, of the to-be-detected face image is determined, wherein T is a positive integer; the face alignment models with the confidence coefficients smaller than a preset threshold value are removed, and the remaining face alignment models continue to be used for carrying out next-stage prediction on the to-be-detected face image; according to the prediction shape, predicted in the Tth stage by the remaining face alignment models in the (T-1) stage, of the to-be-detected face image, the final prediction shape of the to-be-detected face image is determined.

The invention discloses a method and apparatus for predicting spatial distribution position of small-scale heterogeneous geologic body. The method and apparatus are characterized by: using the characteristics of diffraction seismic waves of the small-scale heterogeneous geologic body to extract diffraction wave information from a pre-stack common offset gather and to conduct offset imaging on the diffraction wave information, then conducting attribute analysis which is conducive to prediction of spatial distribution of the small-scale heterogeneous geologic body on a diffraction wave imaging body, finally extracting spatial distribution parameters of the small-scale heterogeneous geologic body, and implementing prediction of the spatial distribution position of the small-scale heterogeneous geologic body in accordance with the extracted spatial distribution parameters. Therefore, the method addresses the prediction of spatial distribution of the small-scale heterogeneous geologic body, and is effective in predicting geologic bodies in areas with complex construction.

Dual-mode compression of images/videos for reliable real-time transmission includes a scalable block-based video compression. The method provides graceful quality degradations in case of channel bandwidth changes. Prediction for encoding is avoided, and the process is less complex with less memory requirements. The method involves processing each block independently with the advantage that errors are not able to propagate beyond current block boundaries. Two different encoding modes are used, natural image coding for acquired images and graphics coding for computer generated images. The proper encoding mode is chosen by comparing performance of the two modes for a particular block to be encoded. An embedded bitstream is generated by the encoder in accordance to the available channel bandwidth.

Multiple words in a text are transformed to unique numerical representations for text mining applications. A web server receives the text, including multiple words in a natural language. A key-word extractor extracts one or more key-words from the received words. A morphologizer morphologizes the extracted key-words based on similarities of fundamental characteristics in the extracted key-words. An analyzer transforms each of the morphologized words to a unique numerical representation such that the transformed unique numerical representation does not result in multiple similar numerical representations.

A method and device for predicting a block of pixels of an image for encoding/decoding same. The method comprises a step of defining a causal vicinity of said block to be predicted, a step 2 of searching for candidate patches during which a set of candidate patches is formed from at least one patch belonging to an image other than the image to which the block to be predicted belongs, each patch being formed from a block and from a causal vicinity of said block, and a step 3 of block prediction during which the block is predicted at least from the block of at least one patch of said set of candidate patches.

A picture encoding and decoding method, a picture encoding and decoding device and a network system are provided, in which, the picture encoding method includes: determining a prediction block used by a picture block according to a division manner of the picture block; determining a corresponding division level in the picture block or the prediction block of transform blocks corresponding to the picture block or the prediction block, in which, the transform blocks corresponding to the picture block or the prediction block include one or more transform blocks; determining a size of the transform blocks corresponding to the picture block or the prediction block according to the division manner of the picture block and the division level; determining identification information for identifying the division level corresponding to the transform blocks corresponding to the picture block or the prediction block; and writing the identification information into a code stream.

A pixel block (300) is compressed by sub-sampling at least a portion of the pixels (310) into subblocks (320, 330). Predictions are determined for the property values of these subblocks (320, 330) by calculating a variance measure based on property values of neighboring pixels (310) /subblocks (320, 330) in two prediction directions in the block (300) relative to a current subblock (320, 330). Ifthe variance is below a threshold, the prediction is calculated based on neighboring property values in both directions. If the measure exceeds the threshold, the neighboring property values in only one of the two predictions directions are used for calculating the prediction. A guiding bit (450) descriptive of the selected direction is also provided. A prediction error is calculated based on theproperty value and the calculated prediction. The compressed block (400) comprises an encoded representation (460) of the prediction error and any guiding bit (470).

The invention provides a program monitoring method and device. The method comprises the following steps of compiling a to-be-monitored program and a monitoring program into class files; by the loadingcharacteristic of a loader, delegating a first type of files and a second type of files to a same top layer for loading; by recompilation, injecting codes of the second type of files into the first type of files; then, obtaining system performance data of a running program corresponding to the first type of files in a loading manner. Without transforming a system is not transformed, an additionalmonitoring assembly is not required to be added. Perception-free, efficient and accurate performance monitoring is achieved, and the performance problem is automatically predicted and analyzed. Intrusion and influence on the system are smaller when efficient performance testing is conducted, the performance index condition of the system and the finer-grained method execution and memory allocationcondition are obtained under the condition that source codes are not modified, and code analysis and data mining can be conducted on obtained data.

A method includes, in a processor, processing a sequence of pre-compiled instructions by an instruction pipeline of the processor. A first block of instructions is identified in the instructions flowing via the pipeline. The first block includes at least first and second conditional branch instructions that conditionally diverge execution of the instructions into a plurality of flow-control traces that differ from one another in multiple instructions and converge at a given instruction. A second block of instructions, which is logically equivalent to the first block but replaces the plurality of flow-control traces by a reduced set of one or more flow-control traces, having fewer flow-control traces than the first block, is created by the processor at runtime. The pipeline is caused to execute the second block instead of the first block.

The embodiment of the invention provides a recommendation method and device based on a knowledge graph, electronic equipment and a storage medium, and the method comprises the steps of obtaining target text information, and searching an entity node matched with the target text information in a pre-constructed knowledge graph; determining a neighbor entity node of the entity node in the knowledge graph, and determining an associated entity node of the neighbor entity node in the knowledge graph; searching a neighbor entity corresponding to the neighbor entity node, and searching an associated entity corresponding to the associated entity node; and recommending the neighbor entity and the associated entity. According to the embodiment of the invention, the knowledge graph is introduced, and recommendation is performed through the knowledge graph, so that a large amount of unknown information can be prevented from being predicted by using less common data among a small amount of users or articles, the possibility of data over-fitting is avoided, and the recommendation accuracy is improved.

A method for encoding a video signal includes determining a dominant prediction direction of a current block using information of a neighboring block; determining a number of intra prediction modes to transmit based on the dominant prediction direction; determining an optimal intra prediction mode based on the number of intra prediction modes; and generating a prediction signal according to the optimal intra prediction mode.

The invention discloses a smart home security communication system and a communication method. The system comprises an instruction initiating terminal, a control center, a smart gateway and smart home equipment. According to the invention, the one-time pad encryption requirement of mass data can be met, the risk that the information of the whole device is stolen is reduced through the double authentication mode of remote access and the double-key XOR encryption and decryption of near-field access, and the security of the smart home system is greatly improved.

The invention provides a trusted computing application method for protecting a nuclear power DCS engineer station. According to the method, the active safety defense of the nuclear power DCS engineerstation is realized through the construction of three levels of safety mechanisms, namely, a trusted root function layer, a trusted system function layer and a trusted service function layer. measurable and controllable active security defense of the whole operation process of the nuclear power DCS engineer station is achieved based on trusted computing, known/unknown viruses or APT can be effectively prevented from being predicted, prevented and controlled through the active security defense, and therefore, the defensive performance and defensive reliability of the nuclear power DCS engineerstation are effectively improved.

The invention discloses a radio frequency device, a server and a radio frequency communication system. The radio frequency device comprises a first key generation unit and a first data processing unit, wherein the first key generation unit is suitable for using real response data of PUF to generate a first key; and the first data processing unit is suitable for using the first key to decrypt the data transmitted by the server or encrypt data to be transmitted to the server. When the above scheme is applied, the communication safety between the radio frequency device and the server can be improved.

The embodiment of the invention provides a model training method, a model training device and an entity extraction method, and relates to the technical field of data processing. The model training method is applied to electronic equipment, and comprises the steps of firstly, obtaining training data and training a preset model according to the training data to obtain a first training model; secondly, obtaining a second training model according to part of parameters of the first training model; and then, training the second training model according to a loss function between the first training model and the second training model to obtain an entity extraction model. Through the setting, the small-scale second training model realizes the function of the first training model, so that the modelscale is reduced on the premise of not reducing the accuracy, and the model prediction and reasoning efficiency is improved.

The invention relates to a method for determining an actual working flow of an oblique-axis pump station. The method comprises the following steps of selecting the positions of upstream and downstreampressure measurement cross sections of a water inlet passage; acquiring geometric parameters of the pressure measurement cross sections; arranging pressure measurement cross section pressure pulsation sensors; acquiring and processing measuring point pulsation data; and calculating equivalent differential pressure of the pressure measurement cross sections and calculating and outputting the flowof the oblique-axis pump station. According to the method for determining the actual working flow of the oblique-axis pump station provided by the invention, through a determined relational expressionbetween equivalent differential pressure on two overflowing cross sections of the water inlet passage and the geometric parameters of the pressure measurement cross sections, an overflowing flow of an oblique-axis pump under the condition of nonuniform distribution of low parameters on the overflowing cross sections of the water inlet passage caused when a pump shaft is obliquely disturbed can beconveniently acquired, and the operation adjustment precision of the working conditions of the oblique-axis pump is improved, so that the oblique-axis pump station is ensured to safely and reliably work.

A big data self-piercing riveting joint key geometrical parameter prediction system comprises a data calling module, a data input module and a data calculation module, and the data calling module is used for calling a pre-trained network model corresponding to a mold type according to the mold type of a to-be-tested joint; the data input module is used for inputting process parameters of the to-be-tested joint into the called pre-trained network model to obtain an opening value and a nail foot self-locking value of a nail foot of the to-be-tested joint; the data calculation module is used forcalculating to obtain key geometrical parameters of the joint to be measured according to the opening value of the nail foot of the joint to be measured; according to the invention, the pre-trained network model is selected according to the type of the mold to calculate the input technological parameters, the nail head height and bottom thickness values of the joint are obtained, and then qualification judgment is conducted on the predicted value result according to the set judgment standard.