44 results about "Dependency relation" patented technology

The application discloses a water conservancy knowledge differentiation retrieval and query decomposition method based on a large language model, and the method comprises the following steps: constructing a standardized retrieval channel for four types of heterogeneous data sources, namely a graph database, a relational database, a document library and a time series database; secondly, establishing a sub-query type classification system for the water conservancy field and forming a template library; thirdly, collecting training data from historical queries and expert samples, and adopting a multi-round labeling mechanism to label query decomposition results, channel mapping relationships and dependency relationships; finally, adopting a low-rank adaptation method to fine-tune the pre-trained large language model, so that the model can intelligently decompose a user natural language query into a sub-query sequence, automatically route to the corresponding retrieval channel and plan an execution sequence. The application realizes intelligent processing of a composite query in the water conservancy field, significantly improves the efficiency and reliability of cross-data source knowledge retrieval, and is suitable for decision support scenes such as flood control scheduling and reservoir group joint scheduling.

Embodiments of the present application disclose a sentence processing method and related device, which are applied to an artificial intelligence scene or a vehicle-mounted scene. For a target sentence and a plurality of first candidate sentences, a sentence similarity between each first candidate sentence and the target sentence is calculated. When it is determined based on the sentence similarity that the plurality of first candidate sentences do not match the target sentence, the plurality of first candidate sentences are segmented to obtain a plurality of second candidate sentences. A plurality of first to-be-fused sentences are determined from the plurality of second candidate sentences according to an importance parameter value of each second candidate sentence for the target sentence. The plurality of first to-be-fused sentences are fused according to a dependency relationship between words in the plurality of first to-be-fused sentences to obtain a fused sentence matched with the target sentence. The method can meet the actual demand for obtaining a matched sentence of the target sentence even when the plurality of first candidate sentences do not match the target sentence, thereby improving the processing effect of application scenarios such as a question and answer system and a search system.

The application relates to the technical field of software test automation, and provides a requirement-driven test case automatic generation method and system based on a semantic test graph, which comprises the following steps: performing multidimensional static analysis on source code, respectively constructing an abstract syntax tree, a control flow graph and a function call graph, and extracting function call relations and data flow dependency relations; constructing a semantic test graph by taking functions as nodes, taking function call relations as main edges and taking data flow dependency relations as auxiliary edges, and aggregating function nodes in the semantic test graph into several function modules; obtaining several test requirement nodes, calculating semantic similarity between the test requirement nodes and the function modules, establishing a mapping relationship from the test requirement nodes to the function modules, and forming a requirement-program structure association graph; and based on the requirement-program structure association graph, generating a test intention for each test requirement node, and converting the test intention into an executable test case. The highly corresponding relationship between the generated test case and the business requirement is ensured.

The application discloses a business auditing method and system architecture based on a large language model and a dynamic DAG arrangement, and relates to the technical field of artificial intelligence and task automation processing. The method is applied to a large language model and includes the following steps: receiving a to-be-audited file and auditing requirement data; performing semantic intention understanding and task disassembly processing on the auditing requirement data to determine atomic tasks; analyzing the dependency relationship between the atomic tasks and generating an initial DAG task graph based on the dependency relationship, wherein the initial DAG task graph includes multiple mutually dependent nodes; determining a current executable node from the initial DAG task graph; executing the current executable node, calling an auditing tool and / or model from a multi-modal tool library to perform business auditing on the to-be-audited file to obtain an auditing result, wherein in the case where the current executable node includes multiple parallel nodes, the multiple parallel nodes are executed in parallel. Through the application, the business auditing efficiency, flexibility and adaptability are effectively improved.

The application provides a cache updating method and system, and relates to the technical field of distribution. The cache updating method comprises the following steps: in response to an invalidation instruction, determining a first node corresponding to first cache data indicated by the invalidation instruction in a dependency graph; the dependency graph is a data structure of data dependency, comprises a plurality of nodes, and the plurality of nodes are connected through edges; the edge represents the dependency relationship between the nodes, and the node represents the cache data; determining a target node having a dependency relationship with the first node; determining the first cache data corresponding to the first node and target cache data corresponding to the target node as invalidation states; and synchronizing the first cache data and the target cache data determined as the invalidation states to a distributed cache.

The application belongs to the technical field of natural language processing, and discloses an intention recognition method, system and device based on a multi-round dialogue dependency relationship graph. By constructing a multi-round dialogue dependency relationship directed graph, the semantic dependency relationship between each round in the dialogue history is explicitly modeled, thereby improving the accuracy of intention recognition in a multi-round dialogue scenario. According to the dependency relationship of a matching node, a history without dependency labels and a dependency history containing dependency labels are constructed; the history without dependency labels and a candidate intention list are input into a large language model for initial intention prediction, and an initial intention is obtained; the dependency history containing dependency labels, a current user speech and the initial intention are input into the large language model for intention recognition and output of a final intention.

The application relates to the technical field of anomaly diagnosis, in particular to an AI application anomaly rapid diagnosis method and system and a storage medium. The method comprises the following steps: extracting corresponding embedded representations between a rule condition expression and a micro-service dependency topology graph, and generating a mapping vector set; triggering a rule adaptation process and updating the mapping vector set when detecting that a dependency relationship change amplitude exceeds a preset amplitude, extracting a cross-service delay propagation link and an abnormal index sequence, and determining a potential fault area set; constructing an abnormal correlation path for the potential fault area to confirm an abnormal root location, forming an adaptation embedded representation mechanism; extracting simulation abnormal scene data and determining an accurate fault area, updating overall corresponding embedded representations, and constructing a final diagnosis rule mapping framework. The application solves the problem that in an AI application environment, service topology dynamic changes cause an abnormal root cause to be difficult to quickly locate, and realizes intelligent identification of an abnormal propagation link and rapid and accurate diagnosis of a fault root.

The application discloses a task processing method and system based on natural language composite intention modeling, comprising: acquiring a natural language composite instruction and performing hierarchical analysis to obtain a plurality of atomic intentions; performing logical concatenation on the atomic intentions to generate dependency relationship triples, combining the atomic intentions to construct a dynamic task execution graph and performing legality verification to generate a standardized execution graph; performing execution component mapping on each atomic intention in the standardized execution graph; controlling each execution component to execute the corresponding atomic intention based on the standardized execution graph, performing parameter transmission according to the dependency relationship triples, and acquiring an execution result; and aggregating the execution result to generate natural language feedback information and outputting the same. Compared with the prior art, the application can overcome the shortcomings of incomplete composite instruction intention analysis, frequent dynamic task logic conflicts and low cross-step parameter transmission accuracy of the traditional method, and effectively improves the instruction analysis accuracy and the logic stability of task arrangement in a complex task scenario.

The application discloses a large language model attention dynamic feature extraction method based on syntax prior constraint and belongs to the natural language processing field. The method is applied to a large language model attention module and comprises the following steps: obtaining a target text and Token of the target text, performing dependency syntax analysis on the target text to determine dependency relation labels of each Token, and calculating a prior vector for syntax perception; obtaining attention column aggregation vectors of each attention head of each layer of the large language model to calculate total attention degrees of each Token by each attention head and alignment scores containing exponential decay penalty terms to quantify calculation effectiveness of each attention head; introducing a temperature coefficient to calculate a refined matrix used as an output of the attention module based on the attention column aggregation vectors of each attention head and the alignment scores. The scheme can improve logicality and structural explainability of an extraction result. In addition, the addition of the exponential decay penalty terms can greatly improve a denoising effect.

The application discloses an error-driven adaptive research process rectification method and system, and relates to the technical field of natural language processing. The method comprises the following steps: decomposing a research task submitted by a user into a plurality of subtasks with a dependency relationship, and generating an initial directed acyclic graph (DAG); executing each subtask according to the DAG, and performing logical verification on an intermediate result to identify an error type; when an error is identified, mapping the error type to a rectification action; adjusting the DAG according to the rectification action, including deleting an error branch, reconstructing a node sequence or inserting a new node; repeatedly executing the process based on the updated DAG until a research completion condition is met, and generating a research report. The application realizes nonlinear re-planning through dynamic adjustment of the DAG, effectively avoids error accumulation, and improves the accuracy and reliability of a complex research task.

The application provides an open source component-oriented code comparison method, system, device and medium, comprising: based on each open source component to be compared, code analysis is performed on the open source component to obtain the dependency relationship between the source codes in the open source component; based on the dependency relationship between the source codes in the open source component, code similarity between each open source component is calculated by using a similarity calculation model to obtain comprehensive similarity information between each open source component; according to the comprehensive similarity information between each open source component, code comparison is performed between each open source component to obtain difference information between each open source component; by using the similarity calculation model based on the dependency relationship between the source codes for code comparison, the similarity analysis is not only limited to the surface syntax dimension, but also extended to the semantic level, thereby facilitating the improvement of the accuracy and comprehensiveness of the code comparison.

The application relates to a compiled graph fusion optimization method, device, equipment, storage medium and computer program product. The method comprises the following steps: acquiring breakpoint information corresponding to a plurality of operation logic segments generated in a front-end tracking process of a computation graph compilation, the breakpoint information comprising operation information triggering a breakpoint, context tensor attributes and a dependency relationship, the dependency relationship being an association relationship between the operation triggering the breakpoint and the context tensor; based on the operation information triggering the breakpoint, the context tensor attributes and the dependency relationship, performing fusion reconstruction on the plurality of operation logic segments to obtain at least one fused logic segment; and performing back-end compilation based on the at least one fused logic segment to obtain compilation information of the computation graph. The method can reduce graph breakpoints.

The application provides a layered multi-task personality detection method based on a large language model. The method first encodes each historical post of a user independently by using a large language model to retain fine-grained personality clues. Secondly, a gating layered context encoder is designed to avoid semantic oversmoothing problems in long text processing through adaptive fusion of local and global information. Finally, a "decoupling-interaction" reasoning paradigm is introduced. Four dimensions of the Myers-Briggs Type Indicator are decoupled through dimensional attention pooling, and the structured dependency relationship between the dimensions is explicitly expressed by using a label graph convolution network. In addition, a class-balanced weighted loss, R-Drop consistency regularization and supervised contrastive learning are combined for joint optimization, which significantly improves the generalization ability and prediction performance of the model in the personality detection task.

This disclosure provides a task processing method, apparatus, electronic device, storage medium, and computer program product, relating to the field of computer technology, specifically large-scale models, natural language understanding, and intelligent agents, and applicable to task processing scenarios. The specific implementation scheme is as follows: a static graph scheduling system determines the task dependency data of the natural language task; the static graph scheduling system processes the natural language task according to a static flowchart; based on the task dependency data and tool description data of tools in a tool set, a dynamic flowchart representing the processing logic of the natural language task is generated; according to the dependency relationships of operators in the dynamic flowchart, the operators in the dynamic flowchart are executed sequentially, generating operator execution results; combining the execution results of each operator, the task processing result of the natural language task is generated. This disclosure can quickly adapt to new tools and complex tool combinations, improving the flexibility of the task processing process and meeting the needs of different business scenarios.

The present application relates to the technical field of natural language processing, in particular to an implicit sentiment analysis method based on syntax enhancement and context awareness, which solves the technical problems that the existing analysis method ignores the difference of sentiment contribution of different syntax relations and easily introduces irrelevant semantic noise, etc., which encodes the word vector of the target sentence and its context sentence to obtain the text semantic feature of the target sentence and the text context feature of the target sentence; a syntax dependency relation weight matrix is constructed to fuse the text semantic feature and the syntax feature of the target sentence to obtain the syntax semantic joint feature; the cosine similarity of the target sentence and its context sentence is calculated and the context features are weighted and aggregated to obtain the aggregated context features; the local to global sentiment reasoning is realized by using the self-attention mechanism, the gated fusion and the hierarchical feature fusion framework based on bidirectional LSTM; the sentence-level global feature is input into the full connection layer to predict the sentiment polarity probability distribution of the target sentence, and finally the sentiment label category of the target sentence is obtained.

This disclosure provides a task processing method, apparatus, electronic device, storage medium, and computer program product, relating to the field of artificial intelligence technology, and particularly to the fields of data processing and intelligent agents. The method includes: using a first intelligent agent to parse at least one of natural language instructions and spreadsheet documents to determine a main task objective; dividing the main task into multiple subtasks with sequential and dependent relationships; for each subtask, using the first intelligent agent to determine an execution plan for the current subtask based on the main task objective and the execution result of the previous subtask; using a second intelligent agent to generate and execute processing code related to the spreadsheet based on the execution plan to obtain the execution result of the current subtask; using a third intelligent agent to match the execution result of the current subtask with the expected objective to determine a valid processing result for the current subtask; and using the first intelligent agent to determine a target processing result based on the valid processing results of each of the multiple subtasks.

The application belongs to the technical field of computer software, and discloses a dynamic binding and analysis method and system for multi-source heterogeneous data in a low-code platform, which comprises the following steps: using a recursive path serialization algorithm, a lexical rule is disassembled into an ordered attribute access sequence, and a business field identifier is extracted based on the ordered attribute access sequence to output a listening queue with a dependency relationship; the listening queue is used as a driving basis to perform multi-dimensional layered retrieval and data aggregation operation, and an evaluation strategy is selected according to the execution result and the complexity of the reference mode, and a memory isolation cloning result is output to complete data binding and analysis processing. The application constructs a self-describing protocol with conflict avoidance capability to solve the problem that static JSON configuration cannot directly express dynamic logic in a low-code environment, and can statically extract field-level fine-grained dependencies and establish accurate responsive listening.

The application relates to the technical field of computers and provides a document automatic generation method and system based on a language generation model, which comprises the following steps: in response to generation request information, creating a task scheduling management unit, defining a plurality of initial tasks and creating corresponding task description bodies, and constructing a task association topology based on data dependency relationships; the task association topology is cyclically scanned, schedulable tasks are identified and submitted to a concurrent execution unit for parallel processing, and a language generation model is called to generate content segments; after target task execution is completed, a new task description body is generated based on the output result, corresponding dependency relationships are established and registered to the task association topology, and result-driven incremental topology growth is realized; after all the tasks are completed, the content segments are synthesized and content integration processing is performed, and a target document is generated. The application realizes efficient concurrent generation and dynamic expansion of the structure of the document, and guarantees context consistency.

The application discloses a kind of based on multi-layer graph embedding atomization component intelligent mapping and process reconstruction method, system, equipment and medium, belong to component reconstruction technical field, comprising: component feature matrix is constructed by multi-source feature extraction, and based on this, multi-layer dependency graph is established to learn component embedding representation;Similarity matching and mapping are carried out using embedding, and the mapping result is optimized in combination with dependency consistency constraint, and then the process dependency topology is constructed and optimized to eliminate redundancy and cycle;Adaptively update dependency graph by detecting historical changes, and introduce expert knowledge constraint to correct and jointly optimize the result, and feed back the revised knowledge to knowledge base.The application realizes unified modeling and deep fusion of component structure, semantics and execution layer dependency relationship by multi-layer graph embedding technology, improves the precision and panoramic understanding ability of component mapping;Ensure the timeliness and accuracy of mapping and reconstruction result.

The present disclosure provides a task execution method and device, electronic equipment and storage medium, relates to the technical field of artificial intelligence, in particular to deep learning, natural language processing and large language model technology. The method comprises: performing semantic analysis on task information, determining a plurality of sub-tasks according to the analysis result; generating a task dependency graph according to the dependency relationship between the plurality of sub-tasks; generating a workflow configuration file according to the information of each sub-task node in the task dependency graph; and executing the plurality of sub-tasks according to the workflow configuration file to obtain the execution result of the task information. The present disclosure improves the logic, stability and efficiency of complex task execution.

The application provides an interface pressure test method, device and equipment, which infers a pressure test scene by using a vertical large model to generate a multi-dimensional scene vector such as load intensity, response delay and bottleneck identification, and then constructs an initial interface dependency graph based on a dependency matrix, calculates a weighted dependency strength by using a multi-type dependency relationship operator combined with a performance index, and performs decoupling processing on the weighted dependency strength based on a dependency type by using a relationship feature decoupling operator to suppress the interference between types, meanwhile, maps the scene vector to a node scene feature component according to the dependency strength and splices the node scene feature component with an interface basic feature to construct a node feature, and finally, based on the node feature, the decoupled dependency relationship and the weighted dependency strength, adopts an improved relationship graph convolution network to perform representation learning to obtain a weighted dependency graph, and performs call chain inference and dynamic pressure propagation path derivation based on the graph, so that more comprehensive and more fine coverage of the new interface pressure test scene is realized.

This invention relates to a multilingual syntax tree generation method based on multi-source syntax guidance and large-scale model collaborative optimization, belonging to the field of natural language processing. The method first generates a dependency syntax tree from the source language text using a traditional syntax analyzer, and performs quality assessment based on the reliability of the root node and dependency relations. For structures with low scores, the Qwen3-8B model is introduced for preliminary correction to improve the quality of the source tree. Subsequently, the segmented target language text and the quality-aligned source language syntax structure are used together as input prompts to guide multiple large-scale models to collaboratively generate the target language syntax tree, achieving cross-language syntax structure transfer. Finally, through consistency integration, rule compliance verification, and format standardization steps, high-quality, uniformly formatted pseudo-syntax tree data is generated. This invention can significantly improve the accuracy of low-resource language syntax analysis and the effectiveness and reliability of data generation standardization in low-resource dependency analysis tasks.

Embodiments of the present application provide an agent scheduling method and device, which belong to the technical field of computers. The method comprises: obtaining an input of a user, the type of the input including text, voice and / or image; pre-processing the input to obtain model input information; based on a large language model, analyzing the model input information to obtain a task instruction, the task instruction including a task type, parameter information and a dependency relationship of at least one subtask; determining an execution order of the subtasks based on the dependency relationship; assigning an agent to each subtask based on the task type of each subtask; and scheduling the assigned agents based on the execution order and the parameter information. The method is used to solve at least one problem existing in the prior art.

The application relates to the technical field of data processing, and provides a kerML meta-model-based architecture framework configuration and generation method. The method comprises the following steps: determining candidate framework elements of an architecture framework by analyzing a kerML meta-model, performing natural language processing on a user demand text, and combining semantic similarity calculation to construct a mapping relationship between elements and intentions; constructing a constraint tensor by integrating instance intention constraints, environment constraints and business constraints, identifying a dominant conflict factor and an influence weight thereof; taking structure parameters and behavior parameters of the architecture framework as decision variables, fusing the dominant conflict factor and the influence weight thereof to construct a fitness function, combining a genetic algorithm iterative optimization to output a configuration scheme set; in an architecture component generation process, verifying interface consistency and constraint compliance of the architecture component, ensuring that the architecture component is qualified and usable, binding the verified architecture component, framework code and a dependency relationship view, and packing to generate an architecture framework instance.

The application provides an intelligent selection method and system for an iterative method for a sparse linear system, wherein after receiving a coefficient matrix and a right end term of a system to be solved, an extended multi-channel coding strategy is used to code the matrix to generate a two-dimensional image, and the right end term is coded to generate a one-dimensional image; subsequently, the features of the matrix and the right end term are extracted through a convolution network respectively, and the obtained feature maps are tokenized; a multi-head self-attention mechanism is used to explicitly capture the internal structure of the matrix, the characteristics of the right end term and the dependency relationship between the two, obtain global tokens, integrate the global tokens into the feature maps, and finally classify the feature maps integrated with the global information to predict the optimal iterative method. Through the whole system representation and global context modeling, the accuracy and robustness of the iterative method selection are effectively improved, and the application is especially suitable for solving complex engineering problems with global dominant characteristics.

The application relates to an event prediction method based on pair event relation learning and multi-view evidence fusion, and belongs to the technical field of natural language processing. The method comprises the following steps: pair event relation learning: a pre-training language model is trained by constructing positive and negative samples in a supervised learning framework; an event relation graph is constructed: the pre-training language model after pair event relation learning is used to deduce the dependency relation of the pair event, the probability output by the model is used as the weight of the edge, and thus the event relation graph is constructed; multi-view evidence learning: quantitative evidence of a candidate event is learned from a text semantic view and a graph structure view; credible evidence fusion: the uncertainty of each view is modeled by using a Dirichlet distribution, and multi-view evidence is dynamically fused by using Dempster-Shafer theory to generate a final prediction result. Under the condition of not depending on an external knowledge base, the application realizes a prediction accuracy of 64.62% on an NYT data set, and exceeds an existing baseline model.

The apparatus employs adaptive machine learning for optimization using dynamically variable local parameters. It consists of a processor and memory. Initially, it access a first dataset corresponding to the first phenomenon and a second dataset corresponding to the second phenomenon. Then, it identifies a dependency relationship between at least one data cluster of the first phenomenon and at least one data cluster of the second phenomenon. Using the at least a processor, modify a processor, an attribute set, as a function of the second data cluster. Further, it optimize a target data cluster, as the function of the first phenomenon. Last, it modify using the at least a processor, the second data cluster as a function of the target data cluster.

The application discloses a retrieval enhancement generation method and system, electronic equipment, medium and program product, and relates to the technical field of artificial intelligence. The method comprises the following steps: calculating the complexity score of a natural language query task; in the case that the natural language query task meets the decomposition condition based on the complexity score, decomposing the natural language query task into at least one subtask, determining the dependency relationship between the at least one subtask; generating a task execution plan of the at least one subtask according to the task dependency relationship between the at least one subtask, determining the target agent corresponding to the at least one subtask, executing the corresponding subtask through the target agent corresponding to the at least one subtask based on the task execution plan, and obtaining the retrieval result of the natural language query task. Thus, the problems that a single retrieval strategy cannot meet diversified query requirements, complex task processing efficiency is low, the advantages of each agent cannot be fully utilized, and the management complexity of task dependency relationship is high are solved.

This invention relates to an aspect-level sentiment analysis method based on the fusion of pre-trained dependency trees and dynamic knowledge graphs, belonging to the field of natural language processing technology. It includes: on the basis of a text sequence representation based on a pre-trained language model, using dependency relation extraction to fine-tune and generate / correct syntactic dependency trees to form a syntactic graph; using a large language model to extract entity / relation triples and constructing a dynamic knowledge graph through relevance filtering and conflict resolution, combining synonym merging and entity alignment to obtain domain knowledge representation; using aspect words as anchors to perform cross-graph alignment of the syntactic graph and knowledge graph, and using a multi-head attention graph convolutional network to achieve cross-graph fusion of syntactic information and external knowledge to obtain aspect-oriented sentiment representation and complete polarity discrimination. This invention utilizes syntactic structure and external knowledge to enhance the recognition ability of implicit aspects and implicit sentiments, improve robustness and interpretability in noisy scenarios, and increase the accuracy of aspect extraction and opinion polarity discrimination.