31 results about "Data-flow analysis" patented technology

A computer-implemented system analyzes program and data flows in a software system comprising code written in multiple programming languages using a generative large language model (LLM) directed by programming-language-specific prompts. The LLM identifies functional components within the code, generating labeled graph nodes that include a node type, a node name, and dependency information. A graph construction computer system processes the labeled graph nodes to generate a directed graph, where nodes represent functional components and directed edges represent dependencies. The system stores the graph in a database and provides a web-based interface for visualization, allowing users to explore, query, and analyze program and data flows across the software system. The system enables automated, language-agnostic dependency mapping, facilitating software analysis, debugging, and modernization.

The invention provides an application packaging processing method, a storage medium and electronic equipment, and is applied to the technical field of software application. According to the method, a full-process closed-loop mechanism of the front stage, the middle stage and the rear stage of packaging is constructed; before packaging, application codes and key parameters are subjected to security detection; in the packaging process, grammar and semantic errors are deeply recognized through a regular engine, an abstract syntax tree and data flow analysis, and a diagnosis report is generated; and after packaging, executing an automatic test by utilizing the matched test case to obtain an application test result. Based on the diagnosis report and the test result, whether repackaging is needed or not is intelligently analyzed, packaging parameters and test cases are adjusted according to the error types and fed back to a safety detection link in a circulating mode, automatic error recognition and self-adaptive optimization are achieved, and therefore the safety and reliability of application packaging are effectively improved, and the development risk is reduced.

The invention provides an evidence-driven large language model MISRA C rule review method and system for aviation safety. According to the method, formalized feature extraction is carried out from four aspects of grammar structure features, semantic structure features, logic analysis features and preprocessing behavior features for target codes according to various rules of MISRA C forcing class rules, and diagnosis information is diagnosed in combination with an industrial-grade compiler; generating a structured evidence set comprising abstract syntax tree node statistical features, code context structure features, control flow graph features, function internal data flow analysis features, symbol and type table features and macro definition analysis features; and based on the structured evidence set and the large language model, executing evidence tracing, rule matching and logical reasoning according to a preset thinking chain process, and outputting a structured judgment result containing illegal rule numbers, evidence description and code positions. According to the method, the false alarm rate can be remarkably reduced while the high rule coverage rate and the detection accuracy are kept.

This disclosure relates generally to method and system to perform interval analysis in source code using functional approach. The method performs data-flow analysis using functional approach to solve infinite-height analyses and properties are validated such as array index within bounds, non-zero division, and preventing arithmetic overflow or underflow on real-life applications. The method receives source code comprising one or more functions to perform a whole program interval analysis over each function using a functional approach to identify range interval of each variable at every program point in the source code. Further, summary for each function is computed which is stored in the form of variable and its corresponding range interval at every exit point of each function. Finally, an incremental interval analysis is performed over each function having edited version change and summary is recomputed for each function impacted by change to optimize overall interval analysis.

The application discloses a kind of based on semantic enhancement's electric power software vulnerability automatic repair method, device and storage medium, the method includes: obtaining the historical source code of target software;According to the abstract syntax tree of preprocessed source code;Based on abstract syntax tree, data flow analysis is carried out to preprocessed source code to construct data propagation chain;The weight score of key symbol is calculated to each function level code segment in preprocessed source code and it is injected as semantic enhancement information into the attention mechanism of model;Additional multilayer perception machine classifier is introduced in the output end of the encoder of model;Preprocessed source code and data propagation chain are input into encoder, and the output of encoder is respectively input into multilayer perception machine classifier and decoder;The decoder generates repair patch according to the output of encoder and multilayer perception machine classifier.The application reduces the computing overhead while ensuring the repair effect, realizes the efficient, accurate and automatic repair of electric power software system vulnerability.

The invention discloses an unmanned aerial vehicle code-oriented static security analysis and vulnerability detection method and system, and relates to the technical field of information security, and the method comprises the steps: carrying out the preprocessing of an unmanned aerial vehicle code, and obtaining standardized code data; performing security vulnerability static detection and performance and complexity analysis on the standardized code data to obtain a vulnerability detection result and a performance analysis result; evaluating and quantifying the obtained vulnerability detection result and the performance analysis result to obtain comprehensive analysis data; and generating a final visual analysis report through the comprehensive analysis data. According to the method, a multi-stage static analysis process is adopted, syntactic analysis, rule detection, data flow analysis and control flow analysis are combined, deep security detection of unmanned aerial vehicle codes is realized, the comprehensiveness and accuracy of vulnerability recognition are improved, two dimensions of security and performance are covered at the same time, and the quality of the unmanned aerial vehicle codes is comprehensively reflected. Quantifiable code quality evaluation is convenient for long-term quality management of an unmanned aerial vehicle project.

This invention proposes a hybrid-mode-based static analysis method for Android software, belonging to the field of malware analysis technology. It primarily addresses the accuracy degradation problem caused by existing Android cross-language static analysis frameworks using function digests for inter-function data flow connections. The main solutions include: function-level data flow analysis of the Android Java layer; lightweight inter-function data dependency analysis of the Android Native layer, generating dynamic function digests for each function; and full-program data flow analysis of the Android Java layer, where data flow connections involving Native function calls are reconstructed using function digests.

The application discloses a kind of data flow analysis methods based on machine learning, comprising the following steps: collecting data flow information, generating data flow characteristic sequence;Build the data flow time series analysis model including feature extraction layer, feature fusion layer and self-learning module;Feature sequence is extracted, and time series embedding vector is generated;Weighted fusion is realized through dynamic attention weight mechanism, and data flow fusion representation vector is generated;The fusion result is calculated, and data flow path, node interaction intensity, prediction bias and abnormal probability information are obtained, and abnormal sample is marked when abnormal probability exceeds threshold value;Model parameters are dynamically adjusted by self-learning module;Based on the updated model, the abnormal risk assessment of real-time data flow is realized.The application fuses long short-term memory network and Transformer structure, has the advantages of strong self-adaptability, high prediction accuracy and excellent real-time performance.

This invention relates to the field of computer technology and provides a method and apparatus for cross-function vector register collaborative allocation for RISC-V vector extensions. The method includes: performing cross-function data flow analysis on the program to identify active vector registers and construct an active information set; establishing LMUL constraint propagation relationships based on this and determining the target LMUL configuration; performing collaborative vector register allocation between the caller and callee based on the active information set and the target LMUL configuration, and establishing a register mapping relationship; performing overflow optimization based on vector characteristics for vector data that cannot be retained; and collaboratively maintaining the vector execution context at function call boundaries. This invention can reduce register reallocation and overflow overhead during cross-function calls, reduce memory access pressure, and improve RISC-V vector register resource utilization and program execution efficiency.

The application designs a detection method for the defect of the UNSAFE code segment in the Rust programming language, first, based on the rustc_lint module, whether the expression in the program language is contained in the unsafe block written by the developer is judged by using HIR; the structure of the expression is analyzed by using HIR; then, the unsafe operation is analyzed and identified by using MIR; finally, the safe expression or statement in the unsafe block written by the developer is reported; the data flow analysis technology for the function parameter is proposed, the alias propagation chain of the function parameter is found out; it is the key technology required for judging whether the function or method is marked with the unsafe keyword, and the target function is tested by combining the fuzz test technology to judge whether the function must be marked with the unsafe keyword.

The invention relates to the technical field of building energy-saving control, and discloses a building energy consumption management method and system. The core lies in that an energy consumption preset constraint condition and a data stream analysis method are invented, energy consumption characteristic analysis is performed on energy consumption data streams, building energy consumption distribution characteristics are identified, the characteristics are combined with environmental data, similarity calibration is performed on a preset energy-saving decision space, a decision trigger interval is determined, and candidate strategies are generated; and the fitness value of each candidate strategy is calculated and screened, so that the feasibility of the scheme is ensured. And further, on the basis of future energy consumption characteristics and environment data obtained through prediction, feedforward compensation optimization is carried out on the screened strategy, and a final optimization scheme is generated and executed. According to the method, through dynamic calibration and prediction optimization, active generation and self-adjustment of an energy-saving strategy are realized, and the intelligent level and energy efficiency of building energy management are improved.

The application discloses a kind of fuzzy testing methods for continuous integration process, it is related to software collaborative development field.The application includes the following steps: in the project of applying continuous integration, difference analysis is carried out to two adjacent submissions, the difference information of both is obtained, and change point is obtained according to difference information, and is stored in change point set;In the process of building project, data flow analysis is carried out according to change point information, and data flow analysis result is obtained;Program is inserted into using data flow analysis result, and the measured program that has been inserted is obtained;The measured program is carried out fuzzy testing, the fitness of seed test case is calculated in the testing process, and corresponding test resource is allocated to it according to fitness.The application more specifically tests the place where change is generated in continuous integration process, and reduces the overhead brought by manual construction test case.

The application discloses a binary program dynamic analysis method based on a process hollowing technology, utilizes a system kernel module to further expand the ability of a traditional binary dynamic plug-in tool when performing fine-grained analysis, constructs a brand-new fine-grained analysis environment in a new hollowing process by combining a code decoupling analysis thought, can reduce the influence on the execution of a target program, improves the application range and analysis reliability of binary analysis, and can reuse the function of an application layer dynamic plug-in tool. The application utilizes the operating system characteristics to realize more flexible and efficient analysis function, gives a hollowing process environment and an analysis task construction method, and can realize fine-grained data flow analysis on special binary programs such as a vulnerability exploitation program and a countermeasure malicious program. The application method can effectively improve the application range of binary program automatic analysis, simplifies the complexity of analysis memory management, and improves the performance and stability of dynamic analysis on binary codes.

This application discloses a data diversion and analysis method, apparatus, and computer-readable storage medium, relating to the field of virtual machine technology. The data diversion and analysis method provided in this application acquires data from a source virtual machine, sets next-hop information according to pre-configured virtual machine parameters, and transmits the data to a preset virtual machine based on the next-hop information; then, the preset virtual machine transmits the data to a destination virtual machine for data analysis. Therefore, when a customer needs to analyze business data, the above solution first transmits the data to a preset virtual machine, and then transmits the data to the destination virtual machine through the preset virtual machine. This three-layer diversion method enables data to be diverted to virtual machines within the same network segment, across network segments, or outside the network for further analysis. This achieves cross-network segment and cross-host virtual machine data analysis, improving data analysis efficiency.

The application provides a kind of redundant code identification method and system based on data flow graph, solve the problem that prior art cannot comprehensively identify multiple types of redundant code. Through the set operation processing of interface access log, full station scanning result and code repository information, a set of abandoned interfaces is identified;Based on static analysis tool and abstract syntax tree, dynamic call detection and data flow analysis are carried out to identify a set of internal useless codes;Then, through the state check and log analysis of consumer running state, script execution record and task scheduling configuration, a set of abandoned scripts and a set of abandoned consumers are identified and merged to form a set of abandoned components;Further, a system call graph is constructed to perform reverse reachability analysis, to identify a set of cascading abandoned codes;Finally, based on the above various sets, a complete redundant code report containing statistical data, detailed list and risk assessment is generated.

The invention discloses a network security policy optimization method and system based on network topology data flow analysis, and belongs to the technical field of network security, and the method comprises the steps: obtaining a configuration state and a data flow transmission demand of each network device in a network topology structure, and generating a network feature set; performing space-time correlation analysis on the cross-security domain data stream, and configuring a strategy priority vector; determining a strategy matching queue configuration parameter and an attack surface risk level through an equipment security level label and a data flow sensitive level identifier in combination with a network feature set and a time delay deviation; generating a control instruction set according to the strategy priority vector, the strategy matching queue configuration parameter and the attack surface risk level; and executing the control instruction set, collecting post attack surface convergence feedback data, generating an equipment policy security thermodynamic diagram, and dynamically adjusting security policy rules of each network equipment in the network topology structure. According to the invention, the cross-security domain data transmission security can be ensured, and the attack surface risk is reduced.

The present application relates to the technical field of image detection, in particular to a model-based X-ray image cigarette detection system, comprising an image acquisition and preprocessing unit, an intelligent recognition and segmentation unit, a data flow analysis and decision unit, a parallel acceleration and resource scheduling unit, a visual interaction and intelligent feedback unit, wherein: the image acquisition and preprocessing unit is used for receiving original ray data output by an X-ray machine, and performing image noise suppression, contrast enhancement and geometric correction by using the parallel computing capability of an image processor. Through exclusive feature extraction, precision-speed adaptive reasoning and model adaptive optimization based on transfer learning, the present application realizes accurate identification, segmentation and quantification of cigarette defects, effectively improves defect detection precision and the adaptation ability of the model to new defects and new specifications of cigarettes, and balances detection precision and speed.

The invention provides a Simulink model code generation method and system based on redundancy elimination. The method comprises the steps that a Simulink model is analyzed to construct a data flow diagram, and the input / output mapping relation of all assemblies is obtained; for a target component in the data flow diagram, determining an accurate output range of the target component by analyzing a data use demand of a subsequent component, and reversely deducing an accurate input range of the target component based on an input / output mapping relation; if the range is reduced, identifying the component as an optimizable component; and for the optimizable component, reorganizing the memory layout of the effective data of the optimizable component, reconstructing an input / output mapping relation, and generating a simplified code which only executes necessary calculation based on the reconstructed input / output mapping relation. Through the global data flow analysis, the CPU load and the memory occupation when the generated code is executed on the embedded system can be obviously reduced, and the execution efficiency and the memory utilization rate of the generated code are improved.

The embodiment of the invention relates to the technical field of software detection, in particular to a data analysis method and device for an application program, equipment and a medium. The method mainly comprises the following steps: performing interprocess data flow analysis on an installation package of an application program to obtain a function call graph, extracting a function sub-graph set called at high frequency in the function call graph, screening the function sub-graph set to obtain a key sub-graph set, and storing the key sub-graph set in a database; the screening standard comprises a correlation standard used for screening sub-graphs containing a vehicle communication output interface function, obtaining a candidate function sub-graph set isomorphic to a topological structure of the key sub-graph set in the function call graph, and simulating code snippets corresponding to the candidate sub-graph set; interaction sequence data between the application program and a preset vehicle simulator is obtained, grammar recovery processing and semantic recovery processing are conducted on the interaction sequence data, and vehicle control instruction information corresponding to the interaction sequence data is obtained. By adopting the method, the data analysis efficiency of the application program can be improved.

This invention provides a multi-stage collaborative intelligent binary vulnerability detection method and system, relating to the fields of software security and vulnerability detection technology. The method includes: disassembling the target binary program, constructing a control flow graph and performing data flow analysis, identifying potential danger points and suspicious code regions using a graph attention network, and generating a static analysis report; using a dynamic binary instrumentation framework combined with a dual-depth Q-network agent to adaptively schedule instrumentation resources and granularity, capturing key runtime states during program execution; using a coverage-guided genetic algorithm combined with a conditional generative adversarial network to generate test cases and perform fuzz testing on the target binary program; and performing correlation analysis to aggregate vulnerability information and generate a comprehensive detection report including vulnerability type, location, and triggering conditions. This invention enables intelligent discovery of binary program vulnerabilities, improving the coverage, accuracy, and generalization ability of vulnerability detection for unknown vulnerability patterns.

Source code in a programming language is received. The source code is converted to a generalized intermediate level representation not specific to any programming language. The source code is converted from the generalized intermediate level representation to a generalized lower level representation adapted to a dataflow analysis portion of static application security testing (SAST). The generalized lower level representation is also not specific to any programming language.

The invention relates to the technical field of artificial intelligence, and discloses a large model execution sequence load generation method, which comprises the following steps: constructing a program control flow diagram corresponding to a large model training program; performing data flow analysis on the program control flow diagram based on the large model execution parameters, and eliminating deterministic condition branches in the program control flow diagram; after a deterministic condition branch in the program control flow diagram is eliminated, statically traversing a large model execution path, determining a target path of a non-deterministic condition branch in the program control flow diagram, and obtaining an execution path sequence of the program control flow diagram; and generating a large model execution sequence load based on the execution path sequence of the program control flow diagram. By combining static analysis and heuristic path selection during compiling, the accuracy of large model execution sequence load generated during coping with complex condition branches is improved. The invention further discloses a device for generating the execution sequence load of the large model, a method for simulation training of the large model and electronic equipment.

The invention discloses a patent correlation evaluation and auxiliary writing method and system, and relates to the technical field of intellectual property, and the method comprises the following steps: S1, capturing a code unit incrementally input by a programmer in real time through a language server protocol or a Java Agent in an integrated development environment, analyzing the code unit based on an abstract syntax tree, and writing the analyzed code unit into a database; constructing a control flow graph and data flow analysis, and generating a structured core technology semantic description; and S2, inputting the semantic description of the core technology into a pre-constructed patent knowledge graph retrieval model. According to the patent correlation evaluation and auxiliary writing method and system, the efficiency and quality of the patent development and protection process are remarkably improved, code input of a programmer in an integrated development environment is captured and analyzed in real time, code core technology semantics are accurately extracted by using an abstract syntax tree and a data flow analysis technology, and the patent correlation evaluation and auxiliary writing efficiency is improved. The patent knowledge graph and the deep learning model are combined, and efficient matching is achieved.

Automated root cause identification using data flow analysis of plural execution traces. A computer system generates data flow dependency graphs from first and second execution traces an entity. These graphs represent input / output data flows of corresponding executions of the entity. The computer system generates topological sortings of those graphs and identifies output pairings across these graphs based on outputs having common labels and topological correspondence. The computer system identifies output pairing(s) that are mismatched as having different values and, for at least one mismatched output pairing, traverses the graphs in order to identify input pairing(s) that are topological root(s) to the mismatched output pairing(s) and that are causal to the mismatch(es). Each input pairing comprises inputs that have a common label, a common topological correspondence, and mismatched values. The computer system returns these input pairings as a root cause for at least one difference between first and second execution traces.

This invention provides a structured text code generation method for PLC honeypots, relating to the field of computer technology. The method includes: constructing a set of protected variables; performing equivalent structural transformations on control blocks and expressions; re-segmenting the code after data flow analysis; and performing statement rearrangement, value modification, and temporary variable renaming on linear statement segments based on data dependency analysis, achieving diverse code perturbations while maintaining semantic consistency. The transformation space is further expanded through loop safety checks and large language model-assisted refinement. Finally, declaration information is supplemented and the output variant L5X file is backfilled. This invention, through semantic awareness and large language model assistance, improves the semantic consistency, formal diversity, and usability of generated code variants.

The invention relates to the technical field of computers, and discloses a program optimization method, a medium, electronic equipment and a computer program product. According to the method, when the source code of the application program is compiled, the unoptimized byte code can be generated. And before the non-optimized byte code is optimized, adding identification information of each closure variable in debugging information corresponding to the non-optimized byte code so as to uniquely indicate one closure variable through one piece of identification information. And then, in the code optimization stage, for each function in the unoptimized bytecode, the data flow relationship between bytecode instructions for operating (such as defining and using) each closure variable can be determined according to the identification information of each closure variable, and the data flow analysis process needs fewer resources and is higher in analysis efficiency. And furthermore, on the basis of the data flow relationship corresponding to each closure variable in the unoptimized bytecode, optimization processing such as redundant closure variable elimination and reading operation can be performed on the unoptimized bytecode, so that an optimized bytecode can be obtained.

Source code in a programming language is received. The source code is converted to a generalized intermediate level representation not specific to any programming language. The source code is converted from the generalized intermediate level representation to a generalized lower level representation adapted to a dataflow analysis portion of static application security testing (SAST). The generalized lower level representation is also not specific to any programming language.

The invention discloses a data flow analysis method based on machine learning, and the method comprises the following steps: collecting data flow information, and generating a data flow feature sequence; constructing a data flow time sequence analysis model comprising a feature extraction layer, a feature fusion layer and a self-learning module; performing feature extraction on the feature sequence to generate a time sequence embedded vector; weighted fusion is realized through a dynamic attention weight mechanism, and a data flow fusion representation vector is generated; calculating a fusion result to obtain a data flow path, node interaction strength, prediction deviation and abnormal probability information, and marking an abnormal sample when the abnormal probability exceeds a threshold value; dynamically adjusting model parameters through a self-learning module; and realizing abnormal risk assessment of real-time data flow based on the updated model. According to the method, the long-short-term memory network and the Transform structure are fused, and the method has the advantages of being high in adaptability, high in prediction precision and excellent in real-time performance.

Type definitions of user-defined types in application program code for which definitions are absent (“unknown types”) are inferred. A static analyzer implements two passes of a fixed-point type inference algorithm. Each pass encompasses a plurality of traversals of the application's control flow to build inferred definitions of unknown types until the inferred definitions are maximally built. To build an inferred definition, based on inferring a variable is an unknown type, the static analyzer infers member variables / functions of the unknown type based on contextual information associated with the variable. Type information of unknown types is propagated along control flow paths. After the first pass terminates, unknown types can be assigned known types based on matching of inferred definitions. Inferred definitions of remaining unknown types are incorporated into the application program code. A second pass of type inferencing and data flow analysis are then performed with the inferred definitions incorporated therein.

The invention relates to a code difference division technology based on data flow analysis, which aims at Java language, designs a static analysis framework taking an abstract syntax tree as granularity, extracts data dependence among syntax tree nodes by using a data flow analysis algorithm, and selects a code difference block in Git as a basic unit to divide the code difference block into the Java language. The relation between syntax tree nodes is improved to be between code snippets, and finally, code difference snippets are divided through a design drawing algorithm to form different code change clusters. According to the method, developers can be assisted in understanding code differences, and the time cost consumed in the code review process is reduced.