61 results about "Data dependency" patented technology

The invention relates to the technical field of distributed computing and computing power scheduling, in particular to a computing power resource fusion method based on a distributed flow pipeline, which comprises the following steps of: disassembling a user computing power request into a flow pipeline unit for packaging computing logic, an input / output interface and a resource demand label; meanwhile, computing power types, real-time load rates, memory occupancy rates, network round-trip delays, geographic positioning and energy consumption data of cloud edge end nodes are collected, a hierarchical topology network framework is constructed based on the collected data, node computing power available values are calculated, an inter-node transmission cost matrix is generated, a fault probability prediction model is constructed, and a fault probability prediction model is constructed. And finally, analyzing a data dependency relationship of the pipeline unit through a four-dimensional joint decision engine, executing dynamic mapping, and preferentially mapping the high-computing-power demand unit to a GPU cluster node, so as to realize non-interruption reconstruction during pipeline topology operation. And the global computing power resource utilization rate, the task operation efficiency and the service stability are improved.

The invention discloses a method for automatically capturing semantic information of a binary code block and a related device, and relates to the technical field of computers.The method comprises the following steps that code disassembling and instruction level analysis are conducted on the binary code block, and a binary stream original instruction sequence is generated; performing semantic intermediate representation conversion and vocabulary overflow standardization on the binary stream original instruction sequence to obtain a binary code standardized instruction sequence; obtaining an instruction data dependency relationship and an instruction control dependency relationship corresponding to the binary code block, and constructing a corresponding binary code program dependency graph; semantic code node embedding is carried out on the binary code program dependency graph to generate a binary code representation vector corresponding to context semantics; and performing automatic semantic capture reasoning on the binary code representation vector corresponding to the context semantics through a pre-trained semantic understanding model to obtain semantic information of the binary code block. According to the method, the semantic information of the binary code block can be efficiently and accurately captured.

The invention provides a lightweight deployment method and system of a large model on edge computing equipment. According to the method, the acceleration capability portrait is constructed by extracting the hardware instruction set architecture type of the target edge device and the number of parallel computing units. Based on the instruction type, the large model weight is grouped, divided and pre-calculated through a unified lookup table vectorization engine, and a pre-calculation vector matched with the target instruction set is generated; and according to the number of the parallel units and the instruction-level parallel capability, compiling the pre-calculation vector to generate an adaptive parallel table look-up instruction block, distributing execution threads with the same number as the parallel units, and eliminating data dependence conflicts. And finally, loading the instruction block to a shared memory area, configuring topological logic of the photoconductive switch matrix based on an instruction type, and dynamically switching a data transmission path in a hardware instruction period. According to the method, the efficient deployment of the large model in the edge equipment and the low-delay reasoning in the resource-constrained environment are realized.

The invention discloses a reservoir level prediction method based on reservoir capacity curve decoupling. The reservoir level prediction method comprises the following steps: 1, forming a sample set by a historical water level and a reservoir capacity sequence; 2, fitting a water level-reservoir capacity curve by using a polynomial; 3, solving an optimal fitting coefficient through a least square method, and establishing a relation curve; 4, calculating the reservoir capacity change in the rainfall period; 5, obtaining a rainfall net contribution amount; 6, calculating a theoretical response water level; 7, collecting historical data and selecting a core predictive factor; and 8, constructing and calibrating a water level prediction model. According to the method, the model which is clear in physical significance, easy and convenient to calculate and high in prediction precision is constructed, reliable technical support is provided for refined scheduling and flood prevention decision making of the reservoir, and therefore the problems that in an existing reservoir water level prediction model, a physical mechanism model is complex in structure, high in data dependence and low in calculation efficiency are solved; and a data-driven model is poor in interpretability, easy to be interfered by data noise and uncertain in generalization ability.

The invention discloses a multi-granularity data cascade updating method and device based on a dynamic topological graph and a medium, and relates to the technical field of data processing. The method maintains a multi-level data dependency model (DAG) in memory. The system recursively updates a target node by monitoring the change of a source node and utilizing an incremental propagation algorithm. For a high-concurrency scene needing isolation verification, the method introduces an overlay shadow topology mechanism: on the premise of not destroying a main graph structure and not copying a total graph, only instantiating a shadow copy of an affected node based on a Context ID, and constructing an overlay Mapping pointing to a father node; when the propagation path is calculated, the numerical value of the shadow node is preferentially read based on the context, and the propagation path of the original node is logically blocked in the current context. The invention further provides a path convergence mechanism and rendering frame synchronization technology based on topology in-degree, and the problems of resource competition, data consistency and front-end rendering flicker under the complex dependence network are effectively solved.

The invention relates to a cross-language software vulnerability detection method and device, and the method comprises the steps: carrying out the analysis of a Joern static analysis pair, carrying out the integration and semantic enhancement of an abstract syntax tree, a control flow graph and a data dependence graph, and obtaining a cross-warehouse heterogeneous code graph; obtaining cross-language intermediate representation based on a compiler framework; after the cross-language intermediate representation and the cross-warehouse heterogeneous code graph are modeled, weighted fusion is carried out through a gated cross attention mechanism, and a multi-modal data set is obtained; carrying out migration training on the multi-modal cross-language vulnerability detection model, and carrying out vulnerability detection on cross-language software to obtain a detection result; through multi-modal data fusion and modeling, in combination with cross-language intermediate representation and a cross-warehouse heterogeneous code graph, the defects of a traditional method in the aspects of cross-language generalization ability and context reasoning ability are effectively overcome; the method has the advantages that the generalization ability of cross-language vulnerability detection is improved, the false alarm rate and the missing report rate are reduced, and the comprehensive utilization effect of global structure information is enhanced.

The invention discloses a distributed AI training-oriented RDMA transmission mode intelligent selection method, and the method comprises the steps: obtaining multi-dimensional data features in a deep learning training process, the multi-dimensional data features comprising a tensor structure feature, a data access mode feature, a data timeliness feature, a data dependency feature and a data change rate feature; identifying a current training stage, wherein the training stage comprises a forward propagation stage, a back propagation stage and a parameter updating stage; based on the multi-dimensional data features and the training stage, an optimal RDMA transmission mode matched with the current training scene is selected through an adaptive decision matrix, and the optimal RDMA transmission mode comprises one or more combinations of RDMA Write operation, RDMA Read operation, Send / Receive operation and Atomic operation; and switching the RDMA transmission mode in real time according to the dynamic change of a network environment, ensuring that a training data flow is not interrupted in a switching process by adopting a progressive flow migration strategy, and optimizing a subsequent mode selection decision through a historical performance learning mechanism.

The invention discloses a Fortran program parallel optimization method based on intelligent dependency analysis. The method comprises the following steps: firstly, constructing a parallel optimization system consisting of a loop extraction module, a loop nested relation analysis module, a semantic analysis module, an intelligent dependency analysis engine, a variable classification module and an instruction generation and injection module; the loop extraction module analyzes a nested relation and a variable action range of loops; the loop nesting relation analysis module determines a nesting relation between loops; the semantic analysis module constructs a row-level data access view; the variable classification module identifies private variables and reduction variables; the intelligent dependency analysis engine executes loop type check, I / O operation check and data dependency check; and the instruction generation and injection module generates a parallelization instruction and inserts the parallelization instruction into the source code to obtain a parallelized program source code. The method can solve the problems that an existing parallelization method is low in cyclic dependency relation recognition accuracy and safety, and parallelization errors cannot be accurately recognized.

The invention discloses a GPU program soft error elasticity prediction method based on an error propagation knowledge graph. The method comprises the steps that static and dynamic information of a program is acquired; single-particle bit flipping fault injection is carried out at the register or memory level, and the occurrence rate of different types of error results after bit flipping occurs at each fault point of the program is obtained; preprocessing data obtained by fault injection, and encoding heuristic features; the method comprises the following steps: pre-training a Word2Vec model based on a program SASS assembly code, and mining semantic knowledge of an instruction to obtain a semantic embedding feature of the instruction; constructing an error propagation knowledge graph by combining the execution condition and the data dependency relationship between the program instructions; and fusing the two heterogeneous feature representations as input, and designing a GPU program soft error elasticity prediction model based on the graph convolutional network. According to the method, heterogeneous feature mining and error propagation relation modeling are carried out on the program, the soft error elasticity of the GPU program fault point can be identified more accurately, and the method has higher soft error prediction rate and lower space-time overhead.

The invention provides a repeated vulnerability detection method based on multi-object type state analysis, and the method comprises the steps: employing a preset vulnerability mode description language, constructing a corresponding vulnerability mode description rule for each vulnerability, and the rule comprises a plurality of variable objects and program statements causing the vulnerabilities; translating the vulnerability mode description rule into a corresponding linear sequential logic formula; converting the formula into an automaton, wherein the automaton comprises an accepting state of the automaton and binding constraints of a plurality of variable objects of the environment in the accepting state; constructing an interprocess control flow diagram for the target program to be detected, wherein the interprocess control flow diagram comprises a plurality of program points; performing multivariable object analysis on the target program based on the graph to obtain an automaton state set of each program point and a data dependency relationship of a plurality of variable objects under each automaton state, and when an automaton state in an accepting state exists and the data dependency relationship of the plurality of variable objects meets the binding constraint, performing the multivariable object analysis on the target program; and detecting that a corresponding vulnerability exists in the target program.

This invention provides a berthing adaptation calculation method and system based on a CAD / CAE composite model, belonging to the field of ship berthing auxiliary decision-making technology. It includes: constructing a loading stability calculation model and a three-dimensional profile model database based on the ship loading manual and an initial model containing geometric information; integrating hull geometry, stability, and hydrodynamic performance data into a unified composite model, and defining data dependencies and update triggering rules; solving heave and surge response characteristics based on a wave model, and calculating limit values ​​in conjunction with wave parameters; calculating the safe range of the air gap height at the platform outlet based on the gangway overlap geometry, combined with limit values, tides, and load changes, and dynamically quantifying the risk; and using real-world data to drive online calibration of the composite model and dynamic simulation of the berthing overlap process. This invention achieves accurate dynamic calculation and risk assessment of the air gap under multi-source data fusion, adapting to the berthing needs of multiple ships, and improving berthing safety and operational continuity.

The invention provides a steel rail corrugation identification method and device based on a neural network fusion model, and relates to the technical field of railway facility detection, and the method comprises the steps: collecting the vertical acceleration data of an axle box of a train; discrete processing is carried out through Fourier transform, low-frequency signals are filtered out, and time domain axle box acceleration data after trend filtering out are reduced through inverse Fourier transform; decomposing the time domain axle box acceleration data through a variational mode decomposition processing method to obtain multi-resolution intrinsic mode function components, and selecting a plurality of components with the highest signal energy ratio as input samples of the model; synchronously inputting the input sample into a convolutional neural network and a bidirectional long-short-term memory network, capturing local features through convolution operation, performing bidirectional time sequence data dependency capturing on axle box acceleration data in combination with a bidirectional mechanism, and combining the convolutional local features with time sequence features through a feature fusion layer of a neural network fusion model to obtain a time sequence data dependency relationship; and obtaining an identification result of the steel rail corrugation section.

The invention relates to the technical field of computers, in particular to a metamorphic relation identification method based on program dependency graph features. The method comprises the following steps: obtaining a control flow graph (CFG) of a target program, and obtaining a control dependency graph according to the control flow graph and a forward dominating tree; obtaining a program dependency graph according to the control dependency graph and the data dependency relationship; according to a program analysis requirement slicing criterion C = (N, V) of a target program, graph traversal is conducted on the program dependency graph, program slicing features are obtained, N represents instructions in the target program, and V represents a variable set; and inputting the program slice features into a target prediction model for identification, and obtaining a prediction result corresponding to the metamorphic relationship of the target program. By the adoption of the method and device, the metamorphic relation obtaining accuracy can be improved, the situation that the metamorphic relation cannot be obtained is reduced, program testing convenience can be improved, and software quality is improved.

Apparatus and methods for detecting non-linear data dependencies in machine learning are provided. The methods may include receiving a data set comprising a plurality of data features, the data set for processing by a machine learning model operating on a machine learning system. The methods may also include plotting the data set in multi-dimensional space and linearizing the plotted, multi-dimensional space by feeding the plotted data to an algorithm programmed to execute the Delaunay triangulation method. The methods may further include receiving from the algorithm the data set linearized in a faceted space and detecting linear correlations in the linearized data set. The methods may additional include identifying a first data feature that has a correlation coefficient with a second data feature that is greater than an accuracy metric and removing the first data feature from the data set to create a reduced data set.

This invention pertains to computer-aided modeling technology, specifically a data-driven, modular CGA (Computer-Aided Architecture) white model generation method and system. The method includes the following steps: defining a CGAModule as a computer-oriented object containing CGA code snippets executed by a CGA engine to generate a 3D model, capable of being mapped to real-world architectural components; constructing a CGA library for managing CGA modeling content, including a CGAModule table, a CGA resource table, and a CGA architectural instance table for archiving; building a visual interface for the CGA library; and using the visual interface to call the CGA library and generate the CGA code for the complete architectural model. This invention generates white model templates by constructing a CGA library and a visual interface, reducing data dependencies and making it suitable for various architectural types and styles.

A microprocessor is disclosed that includes a shared functional unit, a first execution queue, and a second execution queue. The first execution queue includes a plurality of entries, where each entry of the first execution queue includes a first count value that is decremented until the first count value reaches zero. When the first count value reaches zero, the first execution queue dispatches a first type of instruction to the shared functional unit. The second execution queue includes a plurality of entries, where each entry of the second execution queue includes a second count value that is decremented until the second count value reaches zero. When the second count value reaches zero, the second execution queue dispatches a second type of instruction to the shared functional unit. A dispatch unit resolves all data dependencies and resource conflicts such that the first count value and the second count value are preset for the first type of instruction and the second type of instruction to be executed by the shared functional unit at exact times in the future relative to each other.

A resource dependency system and its associated user interfaces, used for tracking data dependencies and data transformations between resources, may display visual node graphs with resources as nodes and the data dependencies and data transformations associated with the columns as edges between the nodes. The nodes representing the resources may be displayed differently based on relevant differences in the resources they represent, which can be set through various selectable criteria and schemes.

The application discloses a code search method and system based on multi-structure graph fusion. An early fusion strategy is adopted to fuse three modal information, i.e. abstract syntax tree, data dependency graph and control flow graph, into a functional program graph as the functional input of the code. Since the edges of the multiple graphs are heterogeneous, a meta-path aggregation graph neural network is adopted to distinguish the control flow, data dependency and syntax tree relationship through different edge type processing mechanisms, extract the code features of the graph, and enhance the understanding of the code semantic structure. The meta-path selection is dynamically adjusted through a graph attention mechanism, further weights are allocated to different nodes, key information is highlighted, the timing information in the code segment is captured, the accuracy of the code representation is improved, and the search capability is enhanced.

The invention relates to the technical field of data analysis, and discloses a pain sensing method and system based on multi-dimensional data analysis, and the method comprises the steps: generating a first pain score based on an existing pain scoring method; the method includes introducing the limb features to generate a second pain score, generating a predicted pain score based on the diagnostic features, and performing a pain intervention based on the predicted pain score and the second pain score. The system corresponds to the method. According to the method, limb features are introduced, and the authenticity of pain characterization is enhanced; by adjusting the weight of the multi-modal data, the data dependence on the face and audio features is reduced, and the evaluation efficiency is improved; a prediction model is constructed in combination with diagnosis data, combination of clinical experience and data analysis is realized, and evaluation reliability is improved; and confidence difference value judgment and dynamic correction are carried out, so that the problem of conflict between multi-modal data and diagnosis results is solved, dynamic adaptive adjustment of pain scores is realized, and the pertinence and effectiveness of a pain intervention scheme are effectively improved.

The present invention proposes a Cholesky decomposition heterogeneous parallel optimization method and system based on the Shenwei architecture, which relates to the field of high-performance computing technology. The method comprises: dividing a symmetric positive definite matrix into sub-blocks based on a distributed parallel allocation scheme and iteratively completing the matrix decomposition; allocating each sub-block to a different process through the MPI programming model, exchanging data between processes through asynchronous communication, and performing coarse-grained task-level parallel acceleration; utilizing the master-slave core acceleration parallel characteristics of the Shenwei architecture to perform two-level parallel acceleration on the four operations in the Cholesky decomposition; wherein, for GEMM and SYRK operations, the column vectors of the matrix are mapped to the slave core array, and the columns are divided according to the number of slave cores; the calculation process is optimized through a double buffering mechanism, vectorized operations and loop unrolling technology to improve parallel efficiency; for the TRSM operation, it is decomposed into multiple TRSV operations, which are allocated to the slave cores for parallel execution, and the data dependency is reduced by loop reading and data broadcasting to achieve efficient parallel computing.

The invention discloses a digital semantic understanding and conversion system and method based on multi-modal feature fusion, the contribution degrees of three types of features including semantics, mathematical and context in different digital type recognition are displayed through a multi-modal feature importance thermodynamic diagram, a macroscopic-microscopic double-layer classification structure is adopted, error propagation is effectively blocked, and the recognition efficiency is improved. Through three-way cooperation of a rule engine, a statistical model and a semantic engine, high-robustness decision making is realized, and an intelligent conversion routing and grammar adaptation mechanism is developed for a digital semantic conversion link. The method can effectively solve the technical bottlenecks of insufficient accuracy, poor expandability, strong training data dependence and the like of a traditional method in complex number type recognition, and is suitable for multiple application scenes such as TTS voice generation, natural language generation, financial document processing, education science and technology and the like; accurate conversion from digital character strings to semantic expressions conforming to human language habits can be achieved.

The application relates to the technical field of human-computer interaction, in particular to a method and device for predicting an embodied intelligent human-computer interaction system based on motion recognition, wherein the method comprises the following steps: collecting three-dimensional coordinates of a plurality of skeleton key points of a human body and time frames corresponding to the coordinates of each skeleton joint to generate initial skeleton data; preprocessing the skeleton data to obtain final skeleton data meeting a condition; performing skeleton coding on the final skeleton data to extract at least one skeleton feature; and training a prediction model based on the at least one skeleton feature, wherein the prediction model comprises a Kronecker product layer, a Kernel Matrix layer and a multilayer perception machine layer to identify a future motion trajectory. Thus, the problem that, in the prior art, due to a large model parameter quantity, strong data dependency and difficulty in intuitively displaying the relationship between variables, the effectiveness and interpretability of the model in actual application are limited, and the trust of a user in an identification result is affected, is solved.

The invention discloses a static taint analysis white-box testing method, device and equipment based on mutual information entropy, and relates to the technical field of test.The method comprises the steps that a taint source and a taint sink are determined, and a program flow chart is constructed; checking the state of the stain sink based on a program dependency graph in the program flow chart, and filtering and optimizing the path of the program to obtain a path simplification graph of the program; performing static taint analysis on the path simplification diagram, performing taint marking on variables with a data dependency relationship according to a taint propagation rule, and calculating a polluted probability of statement variables with a control dependency relationship in combination with a mutual information formula in a quantitative information flow technology so as to realize taint marking; and traversing each dependency set containing the stain mark nodes along the path simplification graph, and generating a stain propagation path graph. According to the method, the taint marking problem under the control dependency relationship can be effectively solved, and the vulnerability injection attack resisting capability and reliability of a program are improved.

The invention belongs to the technical field of artificial intelligence and industrial automation, and discloses an anomaly detection method and system based on a graph neural network and an auto-encoder, and the method comprises the steps: constructing a directed graph to represent a sensor network, carrying out the coding of a device state, fusing the coded device state with node features to reflect physical connection, extracting features, and carrying out the weighted aggregation of node information, the method comprises the following steps: automatically learning weight to highlight contribution of key equipment, reconstructing a sensor state by adopting an auto-encoder, carrying out consistency judgment, outputting abnormal early warning when an error exceeds a threshold value or a prediction deviation is large, and detecting abnormity based on a reconstruction error and a structure inconsistency probability by utilizing an unsupervised learning training model. According to the invention, the accuracy and robustness of the anomaly detection of the switch data of the nuclear power plant are improved, and an unsupervised efficient detection means is provided especially in the environment of complex dependency relationship among multiple data and diversified equipment.

The invention belongs to the technical field of network security. The invention belongs to the technical field of network flow analysis, and particularly discloses a logic vulnerability detection method, system and device for network flow analysis and a medium. Service request response data are captured through network flow, non-request data or static resource data are filtered out by setting a filtering rule, the filtered data are analyzed, and the analyzed data are standardized into structured requests. By automatically identifying API semantics, constructing an API calling chain, a state conversion chain and a data dependence propagation chain, establishing cross-interface automatic identification of data flow and control flow, improving the web system logic vulnerability mining efficiency and coverage, and adding an active alarm function for finding vulnerabilities, the vulnerabilities can be found and solved in time; the method has the advantages that lightweight deployment and low-invasion operation are achieved, an API semantic and state model is automatically constructed, the vulnerability mining intelligent level is improved, the false alarm rate is reduced, and a closed-loop mechanism for discovery, verification and alarm is achieved.

This invention belongs to the field of processor architecture and microarchitecture optimization technology, and relates to a value prediction method for sequential execution processors. It includes: constructing a step value prediction table; recording the most recent true result after the first execution of the target instruction, and forming step information after subsequent executions; generating a predicted value based on the most recent true result and step information when encountering the same static instruction later; writing the predicted value into the corresponding scoreboard entry and propagating it to dependent instructions during the instruction issue phase; writing the true result back to the scoreboard and comparing it with the predicted value after the target instruction is executed; maintaining the existing speculative execution result when the prediction is correct, and pausing front-end instruction fetching, preserving the scoreboard state, and performing a partial reissue recovery on the affected instructions when the prediction is incorrect, before resuming pipeline progression. This invention, while ensuring execution correctness, eliminates some waiting caused by real data dependencies in advance, improving the processor's ability to utilize potential instruction-level parallelism.

This application relates to a hardware acceleration method and apparatus for device fault diagnosis and monitoring based on FPGA. The method loads the parameters of a trained fault diagnosis model into a high-level synthesis tool to construct the basic hardware architecture; it optimizes data flow using row caching and window caching techniques, and achieves parallel data extraction through shift registers, significantly reducing memory accesses and pipeline startup intervals; it achieves computational parallelization through loop structure reconstruction and dependency optimization, eliminates data dependencies using loop swapping techniques, and designs a parallel computing unit array to improve processing efficiency; and it employs double buffering technology to achieve parallel execution of data transmission and computation, further improving system throughput. This method, through a four-step progressive optimization strategy, solves key technical problems of FPGA neural network accelerators, achieves significant performance improvements, provides a complete solution for the hardware deployment of fault diagnosis neural networks, and has significant engineering application value.

The invention discloses a self-adaptive instruction embedding method, which comprises the following steps of: firstly, acquiring an instruction context, extracting control flow and data flow information of a program by using a program analysis tool, then acquiring instruction performance information, acquiring an instruction-level performance label through a performance analysis and simulation tool, and then constructing a semantic model and a performance perception model for training. And finally, establishing a lightweight self-adaptive fusion module, and fusing the output instruction embedding based on self-adaptive weight to obtain the final instruction embedding, wherein the relation between the control dependency and the data dependency of the instruction and the performance characteristics of the instruction are captured. According to the method, three pre-training encoders are designed, context dependence characteristics and performance characteristics of different types of instructions are captured respectively, a constructed lightweight adaptive fusion module dynamically balances the relative importance of three types of embedding in different programs, and the obtained final instruction is embedded for subsequent clustering analysis, so that the accuracy of clustering analysis is improved. And the CPI error rate can be effectively reduced, so that the program can be simulated more efficiently and accurately.

The application provides a programmable neural network inference accelerator and method based on storage access. The neural network inference accelerator comprises a data loading module, an instruction distribution module, a data operation module, a data write-back module, and an on-chip storage unit array. The data loading module, the data operation module, and the data write-back module are in one-way or bidirectional data read-write with the on-chip storage unit array. The instruction distribution module distributes the instructions loaded from outside to the remaining modules, and controls the remaining modules to execute or perform corresponding tasks in parallel, so as to complete neural network deployment. The method can decouple the operation module and the data access module, control the module access to the storage by using the generated instructions in a single instruction multi-cycle forced synchronization manner, so as to transfer the data dependency from the modules to the controlled storage addresses, and enhance the programmability and versatility of the accelerator.