Product
Patsnap Eureka
For R&D, Patsnap Eureka makes reading and utilizing patents & technical documents easy.
Patsnap Eureka AIR
Designed for self-driven R&D workflows. Generate viable solutions, solve complex R&D challenges, empower your innovation with AI.
Patsnap Eureka Materials
Designed for material experts only. Revolutionize your material R&D, from search, analyze, to developing new materials.
Feature
TechResearch
Generate reliable direction feasibility study reports for your R&D in just a few steps.
TechSeek
Discover and master advanced knowledge NOW. Basics, ideas, possibilities, all at once.
TechMind
As an expert in R&D Theories, TechMind can generates customized viable solutions instantly.
TechRisk
Analyze your overall solution with one click, know your potential R&D risks in advance.
TechMonitor
Get weekly tech updates, stay abreast of the latest tech innovations and key insights.
Patents
Literature
Hiro is an intelligent assistant for R&D personnel, combined with Patent DNA, to facilitate innovative research.
326 results about "Computation graph" patented technology
Efficacy Topic
Property
Owner
Technical Advancement
Application Domain
Technology Topic
Technology Field Word
Patent Country/Region
Patent Type
Patent Status
Application Year
Inventor
In simple terms, a computation graph is a DAG in which nodes represent variables (tensors, matrix, scalars, etc.) and edge represent some mathematical operations (for example, summation, multiplication). The computation graph has some leaf variables.
Rapid predictive analysis of very large data sets using the distributed computational graph
InactiveUS20170124464A1Rapid predictive analysisEnsure system stabilityError detection/correctionMultiprogramming arrangementsStreaming dataData stream
A system for predictive analysis of very large data sets using a distributed computational graph has been developed. Data receipt software receives streaming data from one or more sources. In a batch data pathway, data formalization software formats input data for storage. A batch event analysis server inspects stored data for trends, situations, or knowledge. Aggregated data is passed to message handler software. System sanity software receives status information from message handler and optimizes system performance. In the streaming pathway, transformation pipeline software manipulates the data stream, provides results back to the system, receives directives from the system sanity and retrain software.
Owner:QPX LLC
Advanced cybersecurity threat mitigation using behavioral and deep analytics
A system for mitigation of cyberattacks employing an advanced cyber decision platform comprising a time series data store, a directed computational graph module, an action outcome simulation module, and observation and state estimation module, wherein the state of a network is monitored and used to produce a cyber-physical graph representing network resources, simulated network events are produced and monitored, and the network events and their effects are analyzed to produce security recommendations.
Owner:QOMPLX INC
Training neural networks represented as computational graphs
Systems and Methods for training a neural network represented as a computational graph are disclosed. An example method begins with obtaining data representing a computational graph. The computational graph is then augmented to generate a training computational graph for training the neural network using a machine learning training algorithm that includes computing a gradient of an objective function with respect to each of the parameters of the neural network. Augmenting the computational graph includes inserting a plurality of gradient nodes and training edges into the computational graph to generate a backward path through the computational graph that represents operations for computing the gradients of the objective function with respect to the parameters of the neural network. The neural network is trained using the machine learning training algorithm by executing the training computational graph.
Owner:GOOGLE LLC
Platform for autonomous management of risk transfer
A system for autonomous management of risk transfer is provided, comprising a network-connected server configured to provide an interface for a requester to submit a request; an automated underwriting processor configured to: create a contract block by compiling the request into a computational graph-based format, link the contract block to the requester, store the contract block into memory, retrieve a plurality of available underwriting agreements from memory, and create an offer list by perform computational graph operations on the contract block to determine viable risk-transfer agreements; and presenting the offer list to the requester.
Owner:QOMPLX LLC
Neural network compiler architecture and compiling method
ActiveCN110766147AEasy to optimizeImprove targetingNeural architecturesPhysical realisationComputer architectureGranularity
The invention provides a neural network compiler architecture and method. The compiler architecture comprises: a calculation graph construction module which is used for constructing a universal firstintermediate representation based on inputted different types of model files, wherein the first intermediate representation is in a graph form; a calculation graph optimization module which is used for carrying out graph optimization on the first intermediate representation to obtain a second intermediate representation in a graph form; and an instruction generation module which is used for carrying out scheduling optimization on the second intermediate representation to obtain a fine-grained third intermediate representation, and compiling the third intermediate representation into an instruction code executed on the hardware platform based on the hardware platform. The modules in the compiler architecture are matched with various intermediate representations with different granularitiesand attributes, so that various deep learning frameworks and rear-end hardware platforms can be handled with extremely high expandability and compatibility, and efficient and accurate code optimization capability is provided.
Owner:XILINX INC
Knowledge reasoning method based on multi-modal knowledge graph
ActiveCN112288091AIncrease credibilityImprove accuracyKnowledge representationNeural architecturesGraph spectraGraph neural networks
The invention discloses a knowledge reasoning method based on a multi-modal knowledge graph, and aims to enable knowledge reasoning reliability and accuracy to be higher and enable the knowledge reasoning method to have stronger modeling and reasoning capabilities. The method is realized through the following technical scheme: different information is fused based on multi-hop reasoning of a large-scale knowledge base; attribute completion is performed on the attribute missing graph through attribute graph embedding, structured information is extracted from unstructured and semi-structured documents or sentences, and a dynamic heterogeneous graph embedding model is constructed for multi-type characteristics of the multi-modal knowledge graph through heterogeneous graph embedding; feature learning of semi-structured knowledge, structured knowledge and different types of non-structured knowledge is achieved, and multi-modal knowledge graph features are obtained and serve as input for knowledge reasoning based on a graph neural network GNN; an inference path is generated, and a plurality of types of inference paths are constructed; and classification, edge prediction and frequent subgraphs of node types are calculated on the graph, a knowledge reasoning task is generated, and multi-step complex knowledge reasoning is completed.
Owner:10TH RES INST OF CETC
Contextual and risk-based multi-factor authentication
A system for contextual and risk-based multi-factor authentication having a multi-dimensional time series data server configured to monitor and record a network's traffic data and to serve the traffic data to other modules and a directed computation graph module configured to receive network traffic data from the multi-dimensional time series data server, determine a network traffic baseline from the network traffic data, and determine a verification score needed before granting access based at least in part by the network traffic baseline. A plurality of verification methods build up a user's verification score to required level to gain access.
Owner:QPX LLC
Recognizing input gestures using a multi-touch input device, calculated graphs, and a neural network with link weights
Owner:MICROSOFT TECH LICENSING LLC
Transfer learning and domain adaptation using distributable data models
ActiveUS20180197111A1Increase heightImprove distributable modelDatabase distribution/replicationCharacter and pattern recognitionNetwork connectionTheoretical computer science
A system for transfer learning and domain adaptation using distributable data models is provided, comprising a network-connected distributable model configured to serve instances of a plurality of distributable models; and a directed computation graph module configured to receive at least an instance of at least one of the distributable models from the network-connected computing system, create a second dataset from machine learning performed by a transfer engine, train the instance of the distributable model with the second dataset, and generate an update report based at least in part by updates to the instance of the distributable model.
Owner:QPX LLC
Data visualization method and data visualization system based on hierarchical model
ActiveCN105912562AReduce the number of iterationsBeautiful layout effectVisual data miningStructured data browsingGraphicsData visualization
The invention discloses a data visualization method and a data visualization system based on a hierarchical model. The data visualization method comprises the following steps: graph data preparation, graph vertex sampling stratification, sub-graph vertex connection, graph vertex stress calculation, vertex position updating, graph layout recursive calculation, and graph layout hierarchical drawing. The data visualization system comprises a graph data preparation module, a graph vertex sampling stratification module, a sub-graph vertex connection module, a graph vertex stress calculation module, a vertex position updating module, a graph layout recursive calculation module, and a graph layout hierarchical drawing module. Algorithm convergence can be accelerated, the layout can be calculated correctly, and the effect stability is kept. In addition, graph layout of big data can be drawn scientifically, and convenient interactive operation is provided. Therefore, the data visualization method and the data visualization system have the advantage that beautiful layout can be calculated quickly and efficiently and users can be helped to mine the potential knowledge laws.
Owner:SOUTH CHINA UNIV OF TECH
System and method for cybersecurity analysis and score generation for insurance purposes
A system for comprehensive cybersecurity analysis and rating based on heterogeneous data and reconnaissance is provided, comprising a multidimensional time-series data server configured to create a dataset with at least time-series data gathered from passive network reconnaissance of a client; and a cybersecurity scoring engine configured to retrieve the dataset from the multidimensional time-series data server, process the dataset using at least computational graph analysis, and generate an aggregated cybersecurity score based at least on results of processing the dataset.
Owner:QOMPLX INC
Distributable model with biases contained within distributed data
A system for improving a distributable model with biases contained in distributed data is provided, comprising a network-connected distributable model configured to serve instances of a plurality of distributable models; and a directed computation graph module configured to receive at least an instance of at least one of the distributable models from the network-connected computing system, create a cleansed dataset from data stored in the memory based at least in part by biases contained within the data stored in memory, train the instance of the distributable model with the cleansed dataset, and generate an update report based at least in part by updates to the instance of the distributable model.
Owner:QOMPLX LLC
Neural network compiling method for storage and calculation integrated platform
ActiveCN112465108ABalance computing loadImprove parallel efficiencyNeural architecturesPhysical realisationAlgorithmTheoretical computer science
The invention discloses a neural network compiling method for a storage and calculation integrated platform, which relates to the field of storage and calculation integration, and comprises the following steps: analyzing a neural network model, and mapping the neural network model into intermediate representation described by calculation nodes; optimizing a calculation graph; converting into operator-level intermediate representation; carrying out operator task division and binding with a hardware basic unit; performing operator-level optimization, and reducing the number of times of reading discontinuous memories and the number of times of weight mapping. According to the invention, the calculation flow graph and the neural network operator are optimized according to the characteristics of storage and calculation integrated calculation, the expenditure of writing back an intermediate result between graph-level operators is reduced, and the frequency of remapping the weight when the storage and calculation resources are insufficient is reduced.
Owner:SHANGHAI JIAO TONG UNIV
Intermediate representation method and device for neural network model calculation
ActiveCN114186687AEasy to buildLower barriers to useResource allocationNeural architecturesAlgorithmTheoretical computer science
The invention discloses a neural network model calculation-oriented intermediate representation method and device. The method comprises the following steps of S1, analyzing an input model file to obtain topological structure information of a neural network; s2, constructing a logic calculation graph; s21, deducing physical layout information of each operator in the logic calculation graph; s22, deriving the element attribute of each operator in the logic calculation graph; s23, deducing description information of an input and output logic tensor of each operator in the logic calculation graph; s3, constructing a physical calculation graph; s31, generating a physical calculation graph; according to the meta-attribute-based intermediate representation for neural network model calculation disclosed by the invention, data parallelism, model parallelism and pipeline parallelism are originally supported from an operator level. According to the neural network model calculation-oriented intermediate representation method and device disclosed by the invention, the calculation expressions are taken as basic units, the tensors are taken as flowing data in the calculation graph formed by the whole calculation expressions, and the calculation process of the neural network model is realized in a composition manner.
Owner:ZHEJIANG LAB
Systems and methods for neural architecture search
Owner:SWISSCOM AG
Deep learning-oriented data processing method for GPU parallel computing
InactiveCN110941494AIncrease training speedImprove adaptabilityResource allocationNeural learning methodsComputational scienceConcurrent computation
The invention provides a deep learning-oriented data processing method for GPU parallel computing. The method comprises the following steps: firstly, inputting data to model a calculation graph: (1) constructing operation rules of vertexes and edges of a directed graph; (2) defining an execution sequence of operations in the graph by using topological sorting; and (3) updating parameters through the training model, introducing a tensor life cycle, and rewriting a calculation graph based on data operation cost to obtain an optimal operation strategy, mainly comprising the following steps: firstly, modeling the calculation graph based on cost, and redefining an operation function on a CPU; fusing the swap-out operations with the same tensor into a single swap-out operation; and finally, obtaining a traversal sequence by applying a tensor replacement strategy based on calculation and transmission cost. Therefore, the computational graph modeling method based on the formalization rule is constructed. Finally, the extensible neural network and the calculation graph are combined, the training speed of the model can be increased, and the image processing effect is effectively improved.
Owner:HARBIN ENG UNIV
Neural network compiling optimization method and system
The invention provides a neural network compiling optimization method. The method comprises the steps of obtaining a calculation graph with a set data structure according to a to-be-compiled deep learning model; fusing the one or more preprocessing layers into a plurality of fusion layers; obtaining an operator calculation sequence in the fusion layer according to the inter-operator dependency relationship in the fusion layer; obtaining input and output calling times and a splitting strategy of an in-layer operator; obtaining corresponding system overhead values of the plurality of fusion layers on the simulation hardware platform; taking the fusion layer corresponding to the minimum value in the system overhead values of the plurality of fusion layers as the current fusion layer; and compiling the to-be-compiled deep learning model according to the current fusion layer. By fusing multiple layers of operators of the neural network, a calculated intermediate result is stored on a chip instead of being read and written through a memory, so that the memory access requirement can be effectively reduced, and the execution efficiency of the system is improved. Meanwhile, the invention further provides an optimization system for neural network compilation.
Owner:BEIJING TSINGMICRO INTELLIGENT TECH CO LTD
Automated cyber physical threat campaign analysis and attribution
ActiveUS20180159881A1Digital data information retrievalComputer security arrangementsDirect computationData set
A system for automated cyber physical threat campaign analysis and attribution, comprising a multi-dimensional time series and graph hybrid data server, an automated planning service module, and a directed computation graph module. A dataset is gathered from a monitored network and aggregated into a cyber-physical systems graph. Cyberattack simulations on the monitored network are made using exogenously collected data as input. Metrics are generated based on the cyber-physical systems graph and results from the cyberattack simulations, and the generated metrics are used to develop a threat profile.
Owner:QPX LLC
A federated learning model code compiling method and device, electronic equipment and a storage medium
PendingCN112631605AImprove training efficiencyReduce complexityMachine learningCode compilationAlgorithmTheoretical computer science
The invention provides a federated learning model code compiling method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining model information of a federated learning model; constructing a calculation graph corresponding to the model information based on the model information, wherein the calculation graph comprises nodes corresponding to each layer in the federated learning model, and is used for indicating a forward connection relationship and a backward connection relationship between the layers in the federated learning model; obtaining a program code corresponding to each node in the calculation graph; compiling a program code corresponding to each node in the calculation graph according to a forward connection relationship and a backward connection relationship between layers in the federated learning model to obtain an executable training code for training the federated learning model; through the invention, the complexity of training codes of the federated learning model required to be realized by a user can be reduced, and the model training efficiency is improved.
Owner:WEBANK (CHINA)
High-resolution construction land graph spot identification method based on PanTex and linear characteristic
InactiveCN104408463AResolve failureEasy to updateCharacter and pattern recognitionSvm classifierLand use map
Owner:中国土地勘测规划院
Fast and accurate graphlet estimation
ActiveUS20170357905A1Mathematical modelsDigital data information retrievalGraphicsPredictive modelling
Embodiments of the present invention provide a system for fast, accurate, and scalable unbiased graphlet estimation. The system utilizes neighborhood sampling and combinatorial relations to estimate graphlet counts, statistics, and frequency distributions in a small fraction of the computing time of existing systems. The obtained unbiased estimates are highly accurate, and have applications in the analysis, mining, and predictive modeling of massive real-world networks. During operation, the system obtains data indicating vertices and edges of a graph. The system samples a portion of the graph and counts a number of graph features in the sampled portion of the graph. The system then computes an occurrence frequency of a graphlet pattern and a total number of graphlets associated with the graphlet pattern in the graph.
Owner:XEROX CORP
Resource scheduling method, device and system
The invention provides a resource scheduling method. The method comprises the steps of obtaining a job program of a deep learning job, and converting the job program to obtain an intermediate representation of a calculation graph; segmenting the intermediate representation of the calculation graph to obtain a sub-graph set; packaging the sub-graph set to obtain workload mirror images correspondingto various accelerators; and determining a target accelerator from the accelerator cluster according to preset accelerator capability information, the service level condition submitted by the user and the information of the resource pool, and sending a corresponding workload mirror image to the target accelerator. According to the method, unified abstraction is carried out on operation programs of different frameworks by utilizing intermediate representation of the calculation graph; and based on the intermediate representation of the calculation graph, multiple workload mirror images, comprehensive accelerator capability information, service level conditions and information of a resource pool are obtained, a target accelerator is determined, corresponding workload mirror images are allocated to the target accelerator, accelerator resources are reasonably utilized, and the use efficiency is improved. The invention provides a resource scheduling device and system with the above beneficial effects.
Owner:NAT UNIV OF DEFENSE TECH
Drawing method for power distribution wiring diagram of energy utilization information collection system
ActiveCN107450906AAvoid error-prone modification of graphicsImprove loading speedDrawing from basic elementsSemi-structured data retrievalCollection systemArray data structure
The invention relates to a drawing method for a power distribution wiring diagram of an energy utilization information collection system. The method comprises the steps of 1, analyzing an XML document to obtain original data of each graphic primitive, recording graphic data of zooming or rotation operation, performing comparative calculation on the original data of the graphic primitives and the graphic data to obtain a zooming coefficient of the graphic primitives, and performing geometric calculation to obtain drawing center point coordinates of the graphic primitives; 2, based on a canvas API, if a graphic primitive rotation angle is 0, calculating absolute coordinates of the graphic primitives; if the graphic primitive rotation angle is not 0, calculating relative coordinates of the graphic primitives; storing new graphic primitive input parameters obtained by calculation in an object array; and 3, traversing the object array, and drawing graphic primitive object elements one by one. According to the method, the defects of high error probability of graph modification, low page loading speed and the like of an existing drawing method are overcome; the page loading speed is increased; and convenience is brought for system maintenance personnel to draw and modify the graph.
Owner:INTEGRATED ELECTRONICS SYST LAB
Graph neural network model backdoor attack-oriented detection method and device
ActiveCN113297571AProtection securityEasy to detectCharacter and pattern recognitionPlatform integrity maintainanceAlgorithmTheoretical computer science
The invention discloses a graph neural network model backdoor attack-oriented detection method and device, and the method comprises the steps: training a graph neural network model through employing graph data, so as to optimize the parameters of the graph neural network model; inputting the graph data into the parameter-optimized graph neural network model, calculating a loss function corresponding to the graph data, and performing reverse derivation on the loss function relative to an adjacent matrix of the graph data to obtain an importance degree value of each connecting edge to the loss function; extracting sub-graph structures with different connecting edge numbers according to the importance degree values, and dividing the sub-graph structures into a plurality of sub-graph libraries corresponding to the classification labels according to the classification labels; for each sub-graph library, calculating a distribution graph of the sub-graph structures according to the similarity between the sub-graph structures; and analyzing the similarity value in the distribution map corresponding to each sub-map library, and determining whether the map neural network model is attacked or not according to the similarity value. The backdoor attack detection of the graph neural network model is realized, and the security of the model is improved.
Owner:ZHEJIANG UNIV OF TECH
Configurable heterogeneous artificial intelligence processor
PendingCN112463709AImprove scalabilityImprove efficiencyDigital computer detailsProgram controlTerm memoryArtificial intelligence
The embodiment of the invention provides a configurable heterogeneous artificial intelligence processor. The configurable heterogeneous artificial intelligence processor comprises at least two calculation units of different structure types, task queues, storage units and controllers, wherein the task queues, the storage units and the controllers correspond to the calculation units. Each controllerdecomposes a calculation graph of a to-be-processed neural network into a plurality of calculation subtasks and distributes the calculation subtasks to corresponding task queues of all calculation units, and the dependency relationship between all the calculation subtasks is set. According to the set dependency relationship, synchronization among the calculation sub-tasks and controlling access of data related to the calculation sub-tasks is achieved to the storage units and an off-chip memory. According to the technical scheme of the embodiment of the invention, an on-chip heterogeneous formis adopted, and the single controller uniformly schedules and manages the calculation units of each architecture to process different application tasks, so that an artificial intelligence processor can flexibly adapt to different application scenes, the expandability is improved, and the efficiency of processing different tasks is improved.
Owner:SHANGHAI DENGLIN TECH CO LTD
Advanced cybersecurity threat hunting using behavioral and deep analytics
A system for cyber threat hunting employing an advanced cyber decision platform comprising a time series data store, a directed computational graph module, an automated planning service module, and observation and state estimation module, wherein the state of a network is monitored and used to predict network resources that may be vulnerable to a future cyber threat and to produce a cyber-physical graph representing the vulnerable network resources, a human operator is provided with the cyber-physical graph to analyze the data contained therein to initiate an investigation of network resources, and the results of the threat investigation and their effects are analyzed to produce security recommendations.
Owner:QOMPLX LLC
Memory allocation method of neural network
ActiveCN111814971AReduce total memory sizeFully automatedResource allocationMemory adressing/allocation/relocationMemory addressAlgorithm
The invention discloses a memory allocation method of a neural network. A traditional dynamic memory allocation method has great waste, and a manual memory allocation method needs to spend much labortime. The method comprises the steps that firstly, calculation units in a calculation graph are obtained, and all the calculation units are numbered in sequence according to the calculation sequence;obtaining a calculation number set of memory reusable tensors of all calculation units in the model; and determining a final memory allocation mode of the reusable tensors of the memory, and obtainingthe total size of the reusable memory required by the model and the allocated memory address of each reusable tensor of the memory. According to the method, memory fragments generated when the neuralnetwork model applies for and releases the memory can be effectively reduced, the total memory size required by the neural network model is reduced, and the method can be conveniently used in actualengineering.
Owner:HANGZHOU NATCHIP SCI & TECH
Tensor calculation code optimization method and device, equipment and medium
ActiveCN110968321AImprove operational efficiencyEliminate human inputEnergy efficient computingCode compilationTheoretical computer scienceReinforcement learning algorithm
The invention provides a tensor calculation code optimization method and device, electronic equipment and a medium. The method comprises the steps that loop characteristics and calculation graph characteristics of tensor calculation codes are analyzed, corresponding loop information and calculation graph information are obtained, an optimization space is generated according to the loop information, the calculation graph information and a preset optimization method, and each space point in the optimization space represents a preset optimization method combination and parameter selection. Basedon a simulated annealing algorithm and a reinforcement learning algorithm, a target space point is searched and determined in the optimization space; and according to a preset optimization method combination and parameter selection corresponding to the target space point, optimized a tensor calculation code so that automatic optimization of the tensor calculation code can be quickly completed, theoperation efficiency of the tensor calculation code is improved, for programming developers, the human input of a development operator can be avoided, relatively good performance can be obtained, cost can be reduced, and the development efficiency can be improved.
Owner:HANGZHOU WEIMING XINKE TECH CO LTD +1
Acceleration method for exploring optimization space in deep learning compiler
ActiveCN112579063AShort timeReduce overheadNeural architecturesCompiler constructionAlgorithmTheoretical computer science
The invention discloses an acceleration method for exploring an optimization space in a deep learning compiler, and aims to optimize the effect of a neural network through a compiling technology and greatly reduce the time consumed for exploring an operator optimization space by the compiler. The method comprises the steps of firstly abstracting a neural network into a form of a calculation graph;secondly, performing graph optimization on the calculation graph, and defining an optimization space for each operator in the optimized calculation graph; and then, based on an operator containing optimization space information, providing an optimization space similarity calculation method. Finally, an operator state space exploration method based on similarity is provided, operators are clustered based on similarity, full-space exploration is carried out on a core operator in each cluster, other operators of the same type in an optimal scheme of the core operator are explored, and an optimization scheme of each operator of the whole neural network is determined.
Owner:ZHEJIANG LAB
Calculation graph execution method, computer equipment and storage medium
PendingCN111160551AReduce the number of startsReduce the number of memory accessesNeural architecturesPhysical realisationTheoretical computer scienceFusion operator
The embodiment of the invention discloses a calculation graph execution method, computer equipment and a storage medium, and the calculation graph execution method comprises obtaining a binary instruction executed by an artificial intelligence processor corresponding to a calculation graph according to an operation instruction of a fusion operator when a universal processor compiles the calculation graph with the fusion operator.
Owner:SHANGHAI CAMBRICON INFORMATION TECH CO LTD
Who we serve
- R&D Engineer
- R&D Manager
- IP Professional
Why Patsnap Eureka
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com