94 results about "Sorting network" patented technology

A simple format is disclosed and referred to as Elementary Network Description (END). The format can fully describe a large-scale neuronal model and embodiments of software or hardware engines to simulate such a model efficiently. The architecture of such neuromorphic engines is optimal for high-performance parallel processing of spiking networks with spike-timing dependent plasticity. Methods for managing memory in a processing system are described whereby memory can be allocated among a plurality of elements and rules configured for each element such that the parallel execution of the spiking networks is most optimal.

A simple format is disclosed and referred to as Elementary Network Description (END). The format can fully describe a large-scale neuronal model and embodiments of software or hardware engines to simulate such a model efficiently. The architecture of such neuromorphic engines is optimal for high-performance parallel processing of spiking networks with spike-timing dependent plasticity. Neuronal network and methods for operating neuronal networks comprise a plurality of units, where each unit has a memory and a plurality of doublets, each doublet being connected to a pair of the plurality of units. Execution of unit update rules for the plurality of units is order-independent and execution of doublet event rules for the plurality of doublets is order-independent.

A method for improving web search results based on analysis of users' offline activities that are related to their online search. Web search results, including local search results, for a user may be improved by monitoring the actions of users when not interacting with the web search engine. By monitoring “offline” activities (e.g., telephone calls) and matching a person participating in an activity (e.g., the person making a telephone call) to a search engine user, a set of preferences for a given user may be determined, which may be used to select and sort web search results for that user. Preference data for one user may be used to select and sort web search results for other users, including those for which the search engine does not have preference data.

In a network router, a tree structure or a sorting network is used to compare scheduling values and select a packet to be forwarded from an appropriate queue. In the tree structure, each leaf represents the scheduling value of a queue and internal nodes of the structure represent winners in comparisons of scheduling values of sibling nodes of the tree structure. CBR scheduling values may first be compared to select a queue and, if transmission from a CBR queue is not timely, a packet may be selected using WFQ scheduling values. The scheduling values are updated to reflect variable packet lengths and byte stuffing in the prior packet. Scheduling may be performed in multiple stages.

Partial pre-encryption with network-based packet sorting. A video-on-demand (VOD) delivery system for delivering encrypted transport streams to incumbent and overlay set-top boxes utilizes a packet picker / duplicator for sorting selected packets from non-selected packets, duplicating the selected packets, and encrypting one of the pair of duplicated selected packets according to an incumbent encryption scheme. A VOD file server stores the transport stream from the packet picker / duplicator. A network sorter sorts the unencrypted selected packet from the non-selected packets and the encrypted selected packet. The network sorter also sorts the encrypted selected packet from the non-selected packets and encrypts the unencrypted selected packets and the non-selected packets according to an overlay encryption scheme and then sends the transport stream to an overlay set-top box. The network sorter is also responsible for combining the non-selected packets and the incumbent encrypted packets and to send the transport stream to an incumbent set-top box.

An improved system for sorting a data set of 2m keys in a parallel processor environment is disclosed. The system operatively maps a shuffle-exchange network onto a Batcher sorting network in order to produce a monotonically sorted output set. Complete and efficient sorting is achieved by shuffling a control mask applied to multiple compare-exchange elements when the data set is redundantly shuffled. A further improvement is achieved by employing an unshuffle operation to the data set and / or the control mask when the use of an unshuffle operation would be more expedient than use of a shuffle operation.

A simple format is disclosed and referred to as Elementary Network Description (END). The format can fully describe a large-scale neuronal model and embodiments of software or hardware engines to simulate such a model efficiently. The architecture of such neuromorphic engines is optimal for high-performance parallel processing of spiking networks with spike-timing dependent plasticity. Methods for managing memory in a processing system are described whereby memory can be allocated among a plurality of elements and rules configured for each element such that the parallel execution of the spiking networks is most optimal.

The invention relates to a radar signal recovery method based on in-depth study under a complex noise environment, which includes steps of structuring of a radar simulation data set; structuring of anetwork model: recovering a self-adaptive radar signal under the noise environment by a generated confronting network in the in-depth study, and the generated confronting network is composed of a judger and a generator; the judger and the generator are formed by using a thick connection convolution neutral network; training of the generated confronting network: mixing five groups of simulation radar time sequence images under five signal-to-noise ratio environments, generating the training data and training the generated confronting network; recovering the signal of the tested radar image by the generated confronting network.

The invention relates to a browse type service perception analysis method and belongs to the mobile communication field. With the browse type service perception analysis method adopted, massive user browse type service perception data acquired from a terminal side can be effectively utilized to analyze, position and delimit causes in the whole-network level for poor browse type service perception of users, and therefore, related parameters, algorithms, protocols and the like of a network can be optimized, so that the ability of the mobile network to carry browse type services can be improved, and references can be provided for the improvement of the browse type service perception of the users. According to the browse type service perception analysis method, causes for poor browse type service perception are analyzed according to terminal data acquired in actual mobile networks, namely, a plurality of 3G networks of regional and municipal telecommunications companies, and serious problems such as excessively long DNS resolution time and long time delay of the first packet of the Sina website exist in a certain network; and the problems are confirmed by the optimization departments of the telecommunications companies, and the optimization departments of the telecommunications companies carry out corresponding network adjustment, for example, optimize DNS servers and increase local Sina CDN servers, and subsequent tests also verify that the problems are solved, and indexes are improved.

Systems and methods to solve combinatorial problems employ a permutation network which may be modeled after a sorting network where comparators are replaced by switches that controllably determine whether inputs are swapped or are left unchanged at the outputs. A quantum processor may be used to generate permutations by the permutation network by mapping the state of each switch in the network to the state of a respective qubit in the quantum processor. In this way, a quantum computation may explore all possible permutations simultaneously to identify a permutation that satisfies at least one solution criterion. The Travelling Salesman Problem is discussed as an example of a combinatorial problem that may be solved using these systems and methods.

A system and method is provided to accomplish distributed power sequencing function of a large electronics system with minimum number of signals in the sequencing network without compromising the flexibility and expandability. In one embodiment of the invention, the power sequencing function is accomplished with two signals of the sequencing network: power_on / power_off signal and SEQ_LINK signal. The power_on / power_off signal controls whether the sequencing is in power_on mode for turning on power to multiple devices in a predetermined sequence or power_off mode for for turning off power to multiple devices in a reverse sequence. The SEQ_LINK signal controls when the sequence counters, located in each participating device, are allowed to count to the subsequent state. Each sequencing logic circuit of these participating devices responds to a predetermined sequence position to enable the power on or power off of the power supply it controls.

The invention provides a method for acquiring network information, which comprises the following steps of: sorting network sections and acquiring a section list after sorting by counting the network sections accessed by a mobile terminal and access frequency of each section; and acquiring latest network information according to the list and generating a network information list interested by users on the mobile terminal. Through the technical scheme, the latest network information can be provided for the users according to the habits of the users, the users are prevented from logging in a network through a browser at a time and searching interested contents from mass data, so that the time of the users is greatly saved and the operation is simpler, easier, and more convenient and humanized.

In a current anti-virus system, a server can receive virus reports from each user terminal node; when the server receives a large quantity of virus reports from different user terminals, the strategy for evaluating and preferably sorting the virus reports becomes a key point of impacting an entire system to improve the network virus resistance. The invention relates to a method for sorting network virus reports, which is used for anti-virus software capable of being adopted in an Internet or Intranet-based network computing environment. In the method, the virus reports submitted by users are sorted comprehensively according to the number of the virus reports and the confidence level of the users; and units who are used for solving virus problems are allowed to utilize sorted network virus report lists, so that the most valuable virus reports can be preferably processed.

The disclosure provides a distributed cooperative flight control method for a task-oriented formation network of a proximity space platform. The method comprises the following steps: determining a constraint condition of target optimization of a plurality of programming adjacent space platforms; receiving input and output data fed back by the programming adjacent space platforms; and controlling function switching and load distribution of the programming adjacent space platforms based on the constraint condition and the data. The distributed cooperative flight control method for a task-oriented formation network of a proximity space platform provided by the disclosure completes the networking by cooperative control of a plurality of platforms to fully utilize its low cost advantage and achieve the mission goal of network coverage and information transmission in a large area, and performs target optimized decomposition on the plurality of platforms based on the analysis of the constraint condition to realize function switch and effective load distribution control of the programming adjacent space platforms.

The invention discloses a quasi dynamic Huffman hardware encoder and an encoding method; the encoder receives 2n BCD code data each time for encoding, wherein n is a positive integer; the encoder comprises a ranking network module, a tree building module, a tree parsing module, a cascade FIFO, and an output module; the ranking network module and the cascade FIFO are respectively connected with a 4-bit data input port, a start signal port, a clock port and a low level reset port; the ranking network module, the tree building module, the tree parsing module and the output module are connected in sequence; the cascade FIFO is connected with the output module; the output module is provided with a n+1-bit data output port, a n+1-bit data valid port, a data output start signal port and a data output end signal port. The design result shows that when a Nexys4 DDR platform is used, the encoder can work at a frequency above 100MHz; the encoder is high in throughput, low in delay, high in coding efficiency, and simple in structure.

The invention discloses a method and device for realizing importance sorting of network nodes, and the method comprises the steps: obtaining to-be-sorted network nodes of which the number is not greater than N, generating an N * N-dimensional adjacent matrix L containing the weight W and connection edge information among the sorted network nodes, and calculating a static length M for realizing importance sorting of the network nodes according to the adjacent matrix L, and according to the adjacent matrix L and the static length M, calculating an Hermitian matrix A, and according to the value of the sum of the Hermitian matrixes A, outputting a quantum state S containing the importance sorting result of the to-be-sorted network nodes by using a quantum circuit corresponding to an HHL algorithm, the sum of A and S satisfying a linear relationship. By using the embodiment of the invention, index value difference of the to-be-sorted network nodes can be reflected, and the quantum superposition characteristic in the quantum field can be utilized to improve the calculation efficiency.

The invention relates to a method for optimizing a multi-grade light-splitting passive optical network of a distribution communication network. The method is that the network at the current grade bears the gene optimization result of the network at the previous grade through the cascading genetic algorithm. The method specifically comprises the steps of designing a gene code; creating constraint conditions for the gene code, and selecting adaptability functions; performing crossover and variation operation for the gene code on the premise that the constraint conditions for the gene code are met; performing multi-generation iterative crossover and variation operation to obtain the optimal gene code of the PON network at the current level; returning to step (1) to bear the gene code of previous level when the number of the network layers of the multi-grade light-slitting passive optical network is less than n, and updating the adaptability function; finishing the optimization when the number of the network layers of the multi-grade light-slitting passive optical network is more than 1. According to the method, the optimal light splitter and relatively high network star topology in the network are selected from the optimized multi-grade light-splitting network, and therefore, the network communication construction cost is saved; the network planning is constructed into a mathematical model, thus the expandability is improved, and the calculation complexity is reduced.

The invention relates to the field of computer vision, in particular to a multi-target tracking method and system suitable for an embedded terminal. The method comprises the following steps: S1, framing a video to obtain a target image; S2, constructing a multi-target tracking neural network model comprising a detector and a tracker, wherein the detector replaces a feature extraction network of backbone network with ShuffleNetV1; the tracker adopts a Deep-SORT network as a basic network, and uses features extracted by a ShuffleNetV1 network as partial input; S3, performing target detection on the target image to obtain a set of target detection frames; S4, performing state prediction on a tracking object to obtain a target tracking frame; S5, calculating a motion matching degree and a feature matching degree according to results of the tracker and the detector, and constructing a correlation cost matrix; and S6, performing cascade matching on the association cost, determining a tracking result, and realizing a multi-target tracking process. The method solves the problem that an existing multi-target tracking method is limited by hardware equipment and cannot achieve a real-time tracking effect.

The invention discloses a three-dimensional point cloud head posture estimation system and method based on ordered regression and soft labels. The system comprises a feature learning network module which is used for the layered feature extraction of point cloud data; the prediction network module is used for mapping the features obtained by the feature learning network module to a head attitude angle to obtain an angle prediction value, and substituting the angle prediction value and the head attitude angle serving as a label into a first loss function; the sorting network module is used for carrying out dimension division on the head attitude angle to form a plurality of subtasks, generating a soft label according to the relationship between the head attitude angle serving as a label andthe subtasks, carrying out value prediction on the features obtained by the feature learning network module, and substituting the value prediction of the point cloud data and the soft label into a second loss function; network updating module. The loss of the sorting network module and the loss of the prediction network module are combined to introduce the sorting network so as to guide the learning of the prediction network, so that the feature extraction is more accurate, and the precision of the prediction network is improved.

A method that includes identifying a desired signal connectivity through a switch matrix, where the switch matrix includes a plurality of switching elements, and where the switching elements are selectively operable in a plurality of states to provide a plurality of signal paths for routing signals through the switch matrix. The method also includes identifying a sorting network model that corresponds to a topology of the switch matrix, applying a sorting algorithm to the sorting network model, and determining, based on the results of applying the sorting algorithm, operational states of the plurality of switching elements to provide signal paths corresponding to the desired signal connectivity.

The invention provides a method of adopting a firework algorithm to recognize a protein complex. The method includes following steps: converting a protein interaction network into an undirected graph; pre-treating sides and nodes of the protein interaction network; building a dynamic protein interaction network; setting parameters; initializing positions of fireworks; simulating firework explosion to generate spark; selecting part of good points from the spark to serve as fireworks, wherein all fireworks form a class; filtering undesirable classes; outputting a finally-acquired class. The method takes dynamic nature of a protein network and locality and globality of a core-accessory structure inside the protein complex and the protein interaction network into consideration, and the protein complex can be recognized accurately. Simulation experiment results show that performance of indexes like accuracy and recall ratio is excellent. Compared with other clustering methods, the method has the advantages that characteristics of the protein interaction network and the protein complex are combined to recognize the protein complex, and recognition accuracy of the protein complex is improved.

Disclosed herein is a method and system for sorting network videos as a result of an online search. Under the present invention, the system obtains a final sorting factor by calculating weight values of different dimensions, including the release time of a network video, the amount of playbacks of the network video, the classification of the network video and a probability j indicating how often the network video occurs in a collection V. By presenting the videos more needed by the user on the top of a web page, the present invention provides a more accurate display of the search results, thereby improving the user viewing experience.

The invention relates to a system and a method for publishing advertisements on a waterfall-type webpage. The method comprises the following steps: sorting network advertisements which need to be published according to properties; distributing advertisement position carriers at a certain frequency on a webpage presented in a waterfall type; and according to the attributes of the advertisement position carriers, extracting the marks of all advertisement positions; and applying to a background database, and putting advertisements matched according to a certain standard to the advertisement position carriers. The invention also comprises a system which uses the advertisement space carriers for displaying webpage advertisements. The system is divided into an advertisement position carrier distinguishing module, an advertisement position attribute extraction module, a data query module and an advertisement putting module. The invention realizes a novel advertisement display way, and meanwhile, the page loading speed of the webpage can be quickened.

The invention discloses an indeterminate PPI network function module mining method based on fuzzy spectral clustering. The method comprises the following steps: removing repeated interaction and self-interaction in protein-protein interaction data, measuring the interaction of each group by using an edge clustering coefficient, and constructing an indeterminate protein-protein interaction networkchart; improving the similarity measure of the spectral clustering algorithm by using popular distance and the topological properties of the protein network, and further adopting the spectral clustering algorithm to complete preprocessing on protein-protein interaction data; selecting an initial clustering center by using a DPCS strategy, based on the initial clustering center, using a fuzzy C-mean value algorithm to constantly update a clustering center and degree of membership, and then realizing the mining of a protein-protein interaction function module; filtering the mined protein-proteininteraction function module by using the edge expected density of an edge clustering coefficient fused by expected density; and outputting and predicting the mined protein-protein interaction function module.

The invention discloses a pushing and grabbing collaborative sorting network based on double viewing angles and a sorting method and system thereof. The trained pushing and grabbing collaborative sorting network comprises a pushing full convolutional network and a grabbing full convolutional network, and the network is applied to robot pushing and grabbing collaborative sorting. The sorting methodcomprises the following steps of correspondingly acquiring point cloud graphs of an object scene to be sorted from two viewing angles, rotating a top view of the point cloud graphs, correspondingly inputting a plurality of rotating images into the pushing full convolutional network and the grabbing full convolutional network to obtain two thermodynamic graphs with Q values output by the networks,and selecting the thermodynamic diagram with the larger Q value as a final thermodynamic diagram; and according to the pixel point corresponding to the maximum Q value in the thermodynamic diagram and the rotation angle of the rotation image corresponding to the thermodynamic diagram, controlling the robot to execute the sorting action of the network corresponding to the thermodynamic diagram, and then completing sorting. According to the sorting method, double viewing angles are combined with deep Q learning, so that the grabbing success rate is high and the generalization ability is high inthe face of a disordered stacking scene.