24 results about "Syntax error" patented technology

The invention discloses a natural language-to-SQL (Structured Query Language) method and system based on RAG (Random Access Gateway) enhancement, and the method comprises the steps: constructing a query intention graph through dependency syntax analysis, recognizing and complementing semantic missing components, and forming complete semantic representation; vector representation is carried out on the business term segments by adopting vectorization coding, accurate definitions of business terms are obtained from a factory structured knowledge base, and an enhanced context set is formed through expansion retrieval in a low-confidence region; an SQL template mapping network is established based on historical query records, a mapping relation matrix is generated through field candidate matching, mapping conflict positions are identified, and multiple SQL candidate sequences are generated; grammar verification is carried out on the candidate sequence to identify grammar errors, semantic consistency verification is carried out to calculate the intention alignment degree, and an optimal SQL statement is selected through a deviation correction factor; and performing formatting processing and statistical abstract on an execution result, and providing data query capability for scenes such as factory quality management and equipment maintenance.

The invention relates to the technical field of aeronautical text data processing, and discloses an aeronautical text content cleaning and labeling method, system, device and medium wherein the method comprises: noise filtering: identifying and removing noise in an aeronautical text in combination with static cleaning and a general large model; format standardization: converting the aviation text after noise removal into a standardized format text; duplicate removal and error correction: detecting duplicate contents based on a hash algorithm, and correcting spelling errors and grammar errors based on a general large model to obtain an aviation text subjected to duplicate removal and error correction; entity identification: key entities are extracted based on the general large model, and the extracted key entities are labeled; active learning: screening high-value samples in the marked key entities based on an uncertainty query strategy; and dynamic optimization: performing verification and iterative optimization on the marking result in combination with the aviation knowledge base. According to the method, the automation level, the labeling accuracy and the system self-adaptive capability of aviation text processing can be remarkably improved.

The application discloses a lightweight Verilog code generation method based on dual learning, comprising the following steps: adopting a Verilog code corpus to perform domain self-adapting pre-training on a pre-trained language model to obtain a domain-adapted language model; under a dual learning framework, fine-tuning the domain-adapted language model on a paired annotation data set to establish a mapping relationship from a natural language digital circuit hardware design description to a Verilog code, and obtaining a Verilog code generation model; receiving a natural language digital circuit hardware design description input by a user, generating multiple candidate Verilog codes by adopting a probability sampling strategy based on a temperature parameter; verifying the syntax compliance of the multiple candidate Verilog codes, and eliminating the candidate Verilog codes with syntax errors; calculating the confidence of the candidate Verilog codes with syntax compliance, sorting the candidate Verilog codes with syntax compliance according to the confidence, and outputting the Verilog code with the highest confidence as a generation result.

The invention discloses a database SQL statement correction method and device, equipment and a medium, and the method comprises the steps: obtaining an error SQL statement, and determining whether the error SQL statement has a grammar error or not; when it is determined that the wrong SQL statement has a grammar error, correcting a part of the abstract syntax tree corresponding to the wrong SQL statement based on grammar error information corresponding to the wrong SQL statement to obtain a first updated abstract syntax tree; obtaining a database semantic map corresponding to the wrong SQL statement, and determining whether the first updated abstract syntax tree has a semantic error according to the database semantic map; when it is determined that a semantic error exists in the first updated abstract syntax tree, performing semantic correction on the first updated abstract syntax tree according to a database semantic map to obtain a second updated abstract syntax tree; and according to the second updated abstract syntax tree, generating the wrong SQL statement after the grammar and the semanteme are corrected, so that the efficiency and the accuracy of correcting the wrong SQL statement are improved.

The invention relates to a dangerous key scene generation method and system based on big language model retrieval enhancement, and belongs to the technical field of automatic driving simulation testing, and the method comprises the following steps: constructing a scene knowledge base and a trigger condition base based on real traffic accident data; receiving a natural language function test requirement, extracting matched real accident features by using a retrieval enhancement technology, and generating a structured hierarchical scene description; comparing the scene static elements with a pre-labeled map library by adopting a mixed thinking chain compiling strategy with map matching priority; and according to a comparison result, adaptively selecting and multiplexing an existing map file and only generating a dynamic script, or simultaneously generating a static map file and the dynamic script, and finally outputting an executable simulation code. According to the method, a hybrid compiling framework with dynamic and static separation is provided, grammar errors generated by complex road network codes are effectively avoided, and the compilable rate and simulation fidelity of OpenDRIVE and OpenSCENARIO codes are greatly improved.

The invention provides a code generation method and device, equipment, a storage medium and a program product, and can be applied to the technical field of artificial intelligence. The method comprises the following steps: inputting a pseudo code into a code generation model to generate a first candidate code; the error type of the first candidate code is determined, and the error type comprises a grammar error, a code logic error and a code intention error; according to the error type, adopting a corresponding correction strategy to correct the first candidate code to obtain a second candidate code; the second candidate codes are scored based on code quality attributes, and the code quality attributes comprise the conciseness degree of the second candidate codes and the adaptation degree of the second candidate codes and service scenes corresponding to the pseudo codes; and determining a finally generated code according to the scoring result of the second candidate code.

This application relates to the field of automated migration technology for computer software, specifically a declarative method for automated cross-platform UI translation. The method includes: parsing a source UI file, extracting component types, attributes, modifiers, and child node information, and constructing a structured UI tree; mapping the component types, attributes, and modifiers in the UI tree to the basic structure of the target platform ArkUI based on a predefined cross-language component mapping table, generating an ArkUI code skeleton; constructing prompts, inputting the ArkUI code skeleton and prompts into a large language model, and using the large language model to generate preliminary ArkUI code; correcting syntax errors and incompatibilities in the generated preliminary ArkUI code using deterministic transformation rules, and outputting target code that can be directly compiled and run. This invention significantly improves the engineering implementation efficiency of cross-platform UI translation while maintaining high-fidelity visual effects.

An error handling method and apparatus for a video bitstream for determining an error type and a frame type to select a recovery technique. The method receives a video bitstream, detects errors during decoding, and classifies the errors as syntax parsing errors, reference picture data missing or corruption, and decoding errors. The frame type is divided into a reference frame, a non-reference frame or a key frame. Based on these classifications, the method applies appropriate error handling: for partial corruption, generating replacement reference data; for complete reference loss, discarding the frame or the frame group; performing row-level hiding on the reference frame to limit error propagation; performing stripe-level hiding on the non-reference frame to improve visual quality; analyzing key frame errors to ignore slight edge errors; for serious damage, discarding the frame group; for unrecoverable syntax errors, frames are dropped.

A method and an apparatus of error handling for video bitstream decoding determines error types and frame types to select recovery techniques. The method receives a video bitstream, detects errors during decoding, and categorizes errors as syntax parsing errors, missing or corrupted reference picture data, or decoding errors. Frame types are classified as reference frames, non-reference frames, or keyframes. Based on these classifications, the method applies appropriate error handling: generating replacement reference data for partial corruption, dropping frames or frame groups for complete reference loss, performing row-level concealment for reference frames to limit error propagation, performing slice-level concealment for non-reference frames for visual quality, analyzing keyframe errors to ignore minor edge errors or drop frame groups for significant corruption, and dropping frames for unrecoverable syntax errors.

The invention provides a text-to-structured query language generation method and device based on a self-evolution rule base, and the method comprises the steps: obtaining a mixed analysis strategy which comprises a dynamic mapping relation between an intention phrase in a natural language and a structured query language, and carrying out the analysis of a user query based on the analysis strategy, when the initial analysis result is incomplete, the preset language model is called for complementation, grammar errors and field mapping misalignment caused by direct dependence on the model are remarkably reduced through a controllable dynamic mapping relation, and the analysis capability of complex query is ensured; the completion query statement is placed in the isolated security verification environment to be executed, feedback is obtained, and possible error or unauthorized operation is effectively intercepted; and when the execution is successful and the initial analysis result is incomplete, the effective fragment obtained by complementing is automatically converted into a new mapping relationship and added to the dynamic mapping relationship, and repeated manual intervention or model retraining is not needed, so that the maintainability of the text-to-structured query language generation is improved.

The present application belongs to the technical field of conference transcription, and particularly relates to a real-time transcription system device and method based on a multi-language conference system, which comprises a multi-language conference system, a real-time transcription system, an Internet terminal, a self-defined program code and a self-defined extension plug-in, the multi-language conference system is provided with a syntax error compensation function, the multi-language conference system is provided with a paragraph self-definition function, the multi-language conference system is provided with a keyword display function, the multi-language conference system is provided with a multi-language insertion system, one end of the multi-language conference system is connected with the Internet terminal, and one end of the Internet terminal is connected with the real-time transcription device. The real-time transcription system device and method based on the multi-language conference system can complete text editing after transcription, so that the transcribed text is directly usable, and the text has strong watchability during viewing, thereby improving the work efficiency of participants on the transcribed text.

This application discloses a training method and electronic device for a code generation model, relating to the field of computer technology. The method includes: inputting sample prompts into a code generation model to obtain sample code; determining a first reward parameter to characterize execution stability based on the code execution information of the sample code, and parsing the code execution information, sample code, and sample prompts using a reference language model to obtain a second reward parameter to characterize the degree of syntax errors and a third reward parameter to characterize code style; adjusting the model parameters of the code generation model based on the first, second, and third reward parameters; and determining the target code generation model based on the adjusted code generation model when the adjusted model reaches convergence. This solves the technical problem of low training efficiency in code generation models and achieves the technical effect of improving the training efficiency of code generation models.

The invention relates to the technical field of civil engineering and artificial intelligence, in particular to an automatic modeling method and system for a bridge finite element model. Comprising an interaction module, a large model cascade engine, a compiler and a modeling program language grammar error debugging module. The large model cascade engine comprises a natural language pseudo code generation and check module, a modeling program language cue word generation module and a bridge structure parameter generation module; the technical threshold of finite element modeling in the field of civil engineering is greatly reduced, modeling personnel do not need to have a firm computer programming capability, an operable finite element program and a calculation result can be quickly obtained only through input requirements, and the bridge finite element modeling efficiency is remarkably improved.

The application discloses a three-dimensional static layout scheme generation method based on an LLM supporting iterative correction, and comprises the following steps: setting structured data; constructing an LLM, and generating an operation sequence generator and an executable code generator respectively according to the LLM; inputting the structured data into the operation sequence generator for analysis and outputting an operation sequence draft; evaluating the eligibility of the operation sequence draft, and if eligible, determining that the current operation sequence draft is an eligible operation sequence; inputting the eligible operation sequence into the executable code generator for analysis and outputting an executable code draft; and evaluating the eligibility of the executable code draft, and if eligible, determining that the current executable code draft is eligible executable code. The application solves the problem that, in a traditional three-dimensional static layout scheme construction scheme, when structured data is directly converted into executable code with the help of an LLM, the output uncertainty of the LLM leads to logical defects or syntax errors in the generated code.

Embodiments of the present application provide a code highlighting method and device, equipment, storage medium and program product, relating to the field of big data. The method includes obtaining a target code, determining a target language used by the target code; obtaining a syntax highlighting rule of the target language, the syntax highlighting rule including a regular expression matching rule and a syntax scope; performing matching processing on the target code through the regular expression matching rule to obtain a target syntax element; and applying a highlighting style to the target syntax element according to the syntax scope, the syntax scope being used to represent a mapping relationship between the syntax element and the highlighting style. The method can make the target code display a clear and easy-to-read structure, so that the developer can quickly locate and understand the key part of the code, thereby improving the reading efficiency and development efficiency of the code. Meanwhile, the highlighting display can also make the developer more easily find syntax errors in the code.

A data processing system implements receiving a natural language query from a client device, the query including a textual description of one or more actions to be performed using an application on the client device, generating a prompt for a large language model (LLM) based on the natural language query; providing the prompt to the LLM to obtain program code, which when executed, causes the application to perform the one or more actions, the first program code being implemented in a domain specific language (DSL) designed to facilitate program synthesis via the LLM; analyzing a syntax of the first program code to identify syntax errors in the first program code; correcting the syntax errors in the program code to generate corrected program code; and causing the client device to execute the program code to perform the one or more actions in the application on the client device.

Certain embodiments reduce the risks of traditionally programmed algorithms such as syntax errors, unclear logic, and the need for a non-trader programmer to develop the algorithm as specified by a trader by reducing or eliminating the writing of programming code by a user. Certain embodiments provide a design canvas area and blocks for designing an algorithm. Certain embodiments provide for grouping blocks placed in the design canvas area. Certain embodiments provide for virtualized group blocks enabling dynamic instantiation of portions of an algorithm to handle particular discrete events. Certain embodiments provide for operation of some or all portions of an algorithm when a connection between a client device and an algorithm server is broken.

Apparatuses, systems, and techniques to detect errors in content recognized by neural networks. In at least one embodiment, respective document transcriptions of one or more document images are generated using one or more neural networks. The respective document transcriptions may include document content and descriptive information of the document content. An error may be detected in the document content of at least one document transcription of the respective document transcriptions based, at least in part, on a syntax error identified in the descriptive information of the at least one document transcription

The invention discloses a three-dimensional static layout scheme generation method based on LLM supporting iterative correction. The method comprises the following steps: setting structured data; an LLM is constructed, and an operation sequence generator and an executable code generator are generated according to the LLM; inputting the structured data into an operation sequence generator for analysis and outputting an operation sequence draft; evaluating the qualification of the operation sequence draft, and if the operation sequence draft is qualified, judging that the current operation sequence draft is a qualified operation sequence; inputting the qualified operation sequence into an executable code generator for analysis and outputting an executable code draft; and evaluating the qualification of the executable code draft, and if the executable code draft is qualified, judging that the current executable code draft is a qualified executable code. The method solves the problem that in a traditional three-dimensional static layout scheme construction scheme, when structured data is directly converted into executable codes by means of LLM, due to the output uncertainty of the LLM, generated codes have logic defects or grammar errors.

The application discloses a command help information checking method and device, equipment and a readable storage medium, and relates to the technical field of command verification. The method comprises the following steps: identifying a command to be checked, obtaining help information text corresponding to the command to be checked; obtaining a command help information source file corresponding to the command to be checked; comparing the help information text to be checked with the command help information source file, and determining whether there are difference points between the help information text to be checked and the command help information source file; and if there are one or more difference points between the help information text to be checked and the command help information source file, recording the one or more difference points in a checking result file. By implementing the command help information checking method disclosed in the application, the situation that characters are missed and syntax errors occur during manual checking work can be effectively avoided, the checking accuracy of the help information text of the command to be checked is improved, and the command help information text checking efficiency is improved.

Devices, systems, and techniques for detecting errors in content that are recognized by neural networks. In at least one embodiment, respective document transcriptions of one or more document images are generated using one or more neural networks. The respective document transcriptions can include document content and descriptive information of the document content. An error in the document content of at least one document transcription can be detected, at least partially, by means of a syntax error identified in the descriptive information of the at least one document transcription.

The invention provides a compressed code generation method and device, electronic equipment and a readable storage medium. The method comprises the steps that a natural language instruction and context information input by a user are acquired; on the basis of a pre-trained code generation model, under the synergistic effect of an acceleration component and a preset automaton, generating a target compressed code according to the natural language instruction and the context information; the code generation model comprises an encoder and a decoder, and is obtained by performing two-way distillation training on a training sample set formed by samples according to task instruction samples, context samples and original-compression codes; the acceleration component is used for accelerating the reasoning process of the code generation model, and the preset automaton is used for performing grammar constraint verification on phrases gradually generated by the code generation model. According to the method, token redundancy during subsequent model reasoning can be remarkably reduced, the method is not limited by the narrow domain performance of domain specific voice, meanwhile, the problem of grammar errors possibly introduced by code compression is solved, and the consistency of code logic and structure is ensured.

Certain embodiments reduce the risks of traditionally programmed algorithms such as syntax errors, unclear logic, and the need for a non-trader programmer to develop the algorithm as specified by a trader by reducing or eliminating the writing of programming code by a user. Certain embodiments provide building block buttons and an algorithm area to define an algorithm. Certain embodiments provide live evaluation of an expression as the algorithm is being defined. Certain embodiments provide a design canvas area and blocks for designing an algorithm. Certain embodiments provide live feedback for blocks as the algorithm is being designed. Certain embodiments provide for initiating placement of an order to be managed by a selected user-defined trading algorithm from a value axis and for displaying working orders being managed by different user-defined trading algorithms on the value axis. Certain embodiments provide a ranking tool.

The invention relates to a data operation diagnosis method and device, equipment and a medium in the technical field of e-commerce, and the method comprises the steps: obtaining a data operation code provided by a user, and verifying whether the data operation code has a grammar error or not; when the grammar error does not exist in the data operation code, desensitization operation matched with the data type of the service sensitive data is carried out on the service sensitive data in the data operation code, and the desensitized data operation code is obtained; constructing a corresponding diagnosis prompt text based on the desensitized data operation code, and guiding a large language model to diagnose whether the code is correct or not and output a corresponding diagnosis result; and restoring the desensitized service sensitive data mentioned in the modification suggestion in the diagnosis result into the service sensitive data before desensitization, and performing corresponding visual labeling on the data operation code before desensitization according to the error positioning information in the diagnosis result and the restored modification suggestion. According to the invention, low-cost, safe and efficient real-time data operation code diagnosis is realized.