16 results about "Parse tree" patented technology

The invention relates to the technical field of natural language processing and artificial intelligence, and discloses a large-model complex instruction-oriented fine-grained intention recognition method, which comprises the following steps of: constructing a candidate analysis forest of an input instruction, and executing entity alignment of each node and a domain knowledge graph; performing semantic compatibility verification on the candidate dependency trees by utilizing entity relationships in the atlas, and screening an optimal analytic tree in combination with syntactic probability and knowledge consistency scores; if the analytic tree meeting the threshold value does not exist, a semantic conflict edge is positioned, and candidate mounting points are searched by using a map neighborhood relation to reconstruct a dependency structure; and finally, generating standardized data containing a structured analytic path based on the optimal analytic tree, and performing fine adjustment on the large model. According to the method, logic constraint and dynamic repair are carried out on the syntactic structure by introducing the knowledge graph, analysis errors caused by multiple modification or structural ambiguity in a complex instruction are effectively solved, and the accuracy and robustness of intention recognition of a large model in the vertical field are improved.

The application discloses a large language model log analysis method based on relationship inference, comprising: taking a prefix parsing tree as the core, pre-processing the input log, and then traversing the prefix tree to realize template matching. If the matching is successful, the structured log is output, if the matching is unsuccessful, the relationship between the unmatched log and other logs is evaluated from the syntax and semantics. The syntax similarity between the unmatched log and the candidate log set is calculated at the syntax level, the causality strength and the number of shared entities between the unmatched log and the logs in the historical log window are evaluated through LLM at the semantic level, and the semantic correlation log is screened out by combining the semantic composite score; the screened syntax similar and semantic related log is taken as an example to construct a complete prompt containing a task instruction, a demonstration example and a log to be parsed, and a large language model is called to generate a structured template of the unmatched log; the newly generated template is adapted to the prefix parsing tree, the tree structure is perfected through dynamic updating, and more comprehensive template support is provided for subsequent log analysis.

This invention relates to the field of electrical digital data processing technology and discloses a method for detecting log anomalies in power dispatch automation systems. The method includes: constructing a fixed-depth parse tree using historical log data and extracting log templates based on word segmentation similarity; converting power dispatch domain knowledge into vectors and storing them in a knowledge base; retrieving knowledge related to the log templates from the knowledge base, inputting it into a large model for anomaly detection, and labeling the templates; after real-time log preprocessing, inheriting template labels by matching the template library with the parse tree and using an adaptive threshold strategy, triggering online template updates for unmatched logs; aggregating the context and high-frequency parameters of the abnormal logs, constructing dynamic prompt words based on the retrieval results and template feature weights, inputting them into a large model for multi-dimensional root cause analysis, and optimizing the template mechanism based on the prompt words. This method solves the problems of difficult manual maintenance, weak generalization ability, and slow detection speed in existing technologies, achieving efficient detection, high accuracy, and adaptability.

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating an exit scenario output that characterizes an exit scenario for a business enterprise. In one aspect, a method comprises receiving, at an execution engine, data that represents ownership of assets by multiple participants and source code that defines distribution rules of distributing the assets over one or more steps among the multiple participants using a programming language; receiving, at the execution engine, an exit scenario input; automatically parsing, by the execution engine, the distribution rules defined in the source code into a parse tree; and interpreting, by the execution engine, the parse tree to produce an exit scenario output.

A method for training a machine learning model is disclosed, comprising obtaining a training set including multiple sentences and their labeled sets; generating parse trees and label trees for the sentences; determining the total loss of the sentence, which is equal to the weighted sum of the parse tree generation loss and the label tree generation loss; and training the machine learning model using the training set to minimize the total loss. This application also relates to systems, apparatus, and media for training machine learning models and for processing sentences using such machine learning models.

The application discloses a WAF rule evaluation method based on a regular analysis tree, relates to the technical field of WAF rule evaluation, and collects regular expression protection rules of an open source WAF as seeds, analyzes and constructs an expression library based on a regular analysis tree, then performs SQL syntax word recognition on a sent test expression, matches all related expressions in the rule expression library through SQL syntax words, traverses the sent expression request, judges whether the target site WAF protection rule is bypassed according to site feedback information, and further evaluates the rule protection capability of the target WAF system. The application can quickly and accurately deduce the target site WAF protection rule, and can reliably and accurately evaluate the protection capability of the target site WAF system.

One aspect of the present application provides a method for determining check results of scripts, the check results of scripts comprising error detection results in scripts and consistency matching results of scripts, the method comprising the steps of: determining a plurality of data units in each of a plurality of scripts, constructing an analytic tree structure of each script based on the plurality of data units in each script; performing layer sequence traversal detection on each layer in the analytic tree structure of each script by utilizing a preset detection rule to obtain an error detection result in each script; and obtaining at least on the basis of matching partial data subunits in the data unit corresponding to each hierarchy in the parse tree structure of one of the plurality of scripts with partial data subunits in the data unit corresponding to the corresponding hierarchy in the parse tree structure of another of the plurality of scripts, according to the matching result, the consistency matching result of one script and the other script is determined.

The application relates to an ANTLR4-based code generator design method and belongs to the field of code generation. The application generates a lexical analyzer and a syntax analyzer corresponding to an IDL syntax rule file; an IDL file is pretreated, the pretreated IDL file is generated into a syntax analysis tree through the lexical analyzer and the syntax analyzer, and the syntax analysis tree is further generated into structured data; a public interface file for generating C / C++, JAVA files is defined; the C / C++, JAVA public interface file reads a project template file, and structured data in the template file is replaced with structured data in step S2 to generate C / C++, JAVA project files. The application realizes the application of cross-platform and cross-language data distribution services.

A database management system (DBMS) receives an input query and parses the received input query to generate an abstract parse tree. Next, the DBMS traverses the abstract parse tree to detect any parameterized structured query language (SQL) views. If a first parameterized SQL view is detected in the abstract parse tree, the DBMS generates a first view parse tree if a first search of a first cache for the first parameterized SQL view results in a miss. Otherwise, the DBMS retrieves, from the first cache, a previously generated view parse tree if the first search of the first cache results in a hit. Then, the DBMS generates a first query compile tree if a second search of a second cache for the first parameterized SQL view results in a miss. Finally, the DBMS generates and executes a query execution plan based on the first query compile tree.

The application discloses a heuristic online log parsing method and system based on an adaptive deep parsing tree, wherein the method comprises the following steps: constructing a parsing tree, inputting the field number of each log message into different log length matching nodes, and matching the preprocessed log message to the related sub-tree according to the matching result; determining the depth of the current sub-tree according to the length of the preprocessed log message and the preset fitness; matching the several fields of the preprocessed log message to the nodes of the current sub-tree in a head-to-tail alternating manner; after the multi-layer node traversal, reaching the leaf node of the current sub-tree; finding the most matched log group of the preprocessed log message in the leaf node of the current sub-tree; updating the log template and the log ID list of the most matched log group; and obtaining the structured data of the log message set to be parsed after all the log messages are processed.

The invention relates to a data processing method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the steps of obtaining a first query instruction, and performing analysis processing on the first query instruction to obtain analysis tree result data; determining to-be-encrypted field data according to the analysis tree result data; determining a field encryption proportion in the to-be-encrypted field data by using a hierarchical encryption algorithm, and encrypting the to-be-encrypted field data according to the field encryption proportion to obtain an encrypted parse tree; performing serialization operation on the encrypted parse tree, and sending the serialized parse tree to a database server; and obtaining a query result corresponding to the first query instruction returned by the database server, the query result being parse tree result data obtained after the database server deserializes the serialized parse tree, and performing query according to the parse tree result data to obtain result data. By adopting the method, the data security can be provided.

The present application belongs to the technical field of intelligent question and answer, and particularly relates to an experimental question and answer method and system based on a knowledge graph. The method comprises: generating a candidate combined category syntax analysis tree for an input experimental question sentence; when a preset dictionary word is encountered, assigning a temporary combined category by combining a word vector and an adjacent entity type; calculating a syntax probability of the analysis tree and a node density of an entity subgraph in the tree, determining a target syntax analysis tree, and converting the target syntax analysis tree into a logical expression; extracting anchor entities and semantic relationships, determining a search depth according to the subgraph density, and constructing an initial graph query sentence by using bidirectional breadth-first search; if the query has no result, differentiating the query reconstruction according to the density threshold, replacing only the semantic relationship with a parent class when the density is high, or considering both the parent class and the sibling class relationship, and outputting a subgraph or an entity obtained by the query as an answer. The present application can improve the adaptability, accuracy and fault tolerance of the experimental question and answer system, and balance the query efficiency and recall rate.

The embodiment of the invention provides a quality evaluation method and device for SQL statements and a storage medium. The quality evaluation method and device are used for accurately evaluating the quality of the SQL statements. The method comprises the steps that a first instruction is received, the first instruction comprises an execution plan text, the execution plan text comprises a plurality of plan segments, and the plan segments comprise execution plan information of a first SQL statement on execution nodes; the first instruction is used for indicating to analyze the execution plan text and returning an index value of at least one first index based on an analysis result; in response to the first instruction, inputting the execution plan text to a first parser, generating a first parse tree, extracting an index value of at least one first index from the first parse tree, and sending the index value to an interaction device, the method and the device are used for determining the quality score corresponding to the first SQL statement by the interaction device, so that the quality of the SQL statement can be accurately evaluated.

The invention discloses an aerospace fault diagnosis method based on generative intelligence, and belongs to the technical field of aerospace fault diagnosis. The spaceflight fault diagnosis technology based on generative intelligence provided by the invention can automatically extract fault diagnosis knowledge in documents such as FMEA, a plan disposal manual and a historical return-to-zero report, and construct a spaceflight troubleshooting knowledge graph; when an exception occurs, a user can quickly retrieve fault knowledge in a natural language description mode and associate the fault knowledge to form a fault cause analysis tree; after the fault reason analysis tree is confirmed, the system can automatically construct an intelligent agent to execute a fault reason troubleshooting process, assists in completing fault reason screening or positioning, supports to provide a fault troubleshooting suggestion for a user in a natural language interaction mode, and improves the user experience after the fault reason is finally positioned. The fault diagnosis report can be automatically generated, and the knowledge graph can be updated.

The application relates to a cloud service source code-based multi-mode architecture reverse analysis method, which comprises the following steps: acquiring program source code on a warehouse on a cloud server through a tool, and then completing reverse analysis of the architecture of the whole program through source code analysis. The method can call an analysis tool, acquire an abstract syntax tree of all source code files, construct a program analysis tree of a source code project according to abstract syntax tree information, analyze element (file, class, function and variable) coupling information of the program analysis tree, construct a dependency relationship between tree nodes, integrate element coupling information of the program analysis tree, and construct a dependency relationship graph of all source code files. The generated source code file dependency graph can be used to aggregate a dependency relationship graph of a source code module and a relationship dependency graph of a component according to a classical mode, a Maven project single-module mode and a Maven project multi-module mode.

A database management system (DBMS) receives an input query and parses the received input query to generate an abstract parse tree. Next, the DBMS traverses the abstract parse tree to detect any parameterized structured query language (SQL) views. If a first parameterized SQL view is detected in the abstract parse tree, the DBMS generates a first view parse tree if a first search of a first cache for the first parameterized SQL view results in a miss. Otherwise, the DBMS retrieves, from the first cache, a previously generated view parse tree if the first search of the first cache results in a hit. Then, the DBMS generates a first query compile tree if a second search of a second cache for the first parameterized SQL view results in a miss. Finally, the DBMS generates and executes a query execution plan based on the first query compile tree.