142 results about "Readability" patented technology

The invention provides a front-end automatic testing method of a large language model based on LLM, and belongs to the technical field of front-end automatic testing. The method comprises five core steps of test intention recognition and semantic analysis, page structure analysis and target element positioning, automatic generation of a test operation sequence and assertion logic, test script execution and dynamic regression verification, and test feedback analysis and continuous training. The strong semantic understanding and context reasoning capabilities of the LLM are utilized to perform semantic analysis and simulation on a front-end page structure, a user behavior process and business logic, so that automatic generation and maintenance of a test script are realized. According to the method, the test coverage rate and the generation efficiency are greatly improved, the maintenance cost is remarkably reduced, the readability and the intelligence of the test case are enhanced, and the method is suitable for the modern Web front-end application test of rapid iteration.

The invention discloses a multi-language text self-adaptive configuration method and electronic equipment, and relates to the technical field of computers, and the method comprises the following steps: obtaining a source text set in response to a translation instruction received when a page runs, and translating the source text set according to a target language identifier and a constraint strategy to obtain a translated text set; according to a constraint strategy, carrying out adaptability detection on each translation in the translation set to obtain an adaptive set and a non-adaptive set; performing semantic rewriting on each non-adaptive translation in the non-adaptive set to obtain a target candidate set; and aggregating the adaptation set and the target candidate set in the same rendering frame, and rendering and displaying the adaptation set and the target candidate set in batches, so that the problems of lack of multi-language dynamic adaptation capability and insufficient semantic equivalent compression are solved, the accumulated layout offset is remarkably reduced on the premise of ensuring semantic integrity and readability, and the layout efficiency is improved. And the page stability and the user experience are improved.

The invention discloses an industrial defect image-text joint detection method and related equipment. The method comprises the following steps: acquiring an industrial product image and extracting multi-scale visual features; obtaining text information related to the image and extracting text features; the image and text features are input into a large language model with a cross-modal attention mechanism for fusion reasoning, and defect semantic features are generated; outputting a pixel-level defect mask through a segmentation decoding module based on the features; and finally, jointly generating defect text description and superposing the defect text description with the original image to form an image-text detection result to be output. The method makes full use of complementarity of visual perception and text knowledge, has the advantages of low sample dependence, high robustness and human readability of output results, is suitable for quality monitoring and defect analysis scenes in the industrial manufacturing process, and has good practicability and popularization value.

The invention belongs to the field of contract document translation, and provides a contract document translation method and system based on a large model, and the method comprises the steps: obtaining a target contract document, recognizing a picture in the target contract document, converting the picture into a text, carrying out the recognition of the text in the target contract document through taking the recognized line feed as a mark of a block, and taking a page number as a maximum unit, the method comprises the following steps of: performing preliminary blocking to form a plurality of text blocks, identifying character features in each block during preliminary blocking, correcting the text blocks according to the identified character features, merging the text blocks belonging to the same natural segment, and splitting the text blocks of incoherent texts; and translating the corrected text blocks by using a pre-trained large model, and writing the translated text blocks into a new file in a form conforming to the character features corresponding to the text blocks to obtain a translated contract document. According to the method, the professionality, correctness and readability of translated documents are improved.

The invention discloses a code annotation generation method and device, electronic equipment, a storage medium and a product, which are applied to the technical field of computers and are used for solving the problem of low efficiency caused by manual annotation writing in the prior art and effectively ensuring the accuracy and practicability of annotation so as to improve the readability of codes. Determining an annotation prompting rule; the annotation prompting rule comprises a corresponding relation between a code statement type and a code annotation type; analyzing the target program code by utilizing a large language model, and determining a to-be-annotated code snippet in the target program code and semantic information and a statement type of the to-be-annotated code snippet; determining an annotation type corresponding to the to-be-annotated code snippet according to an annotation prompt rule and the statement type of the to-be-annotated code snippet; and generating a target annotation text corresponding to the to-be-annotated code snippet according to the annotation type and semantic information corresponding to the to-be-annotated code snippet, and inserting the target annotation text into an annotation position corresponding to the to-be-annotated code snippet.

A method for extracting key phrases based on extracting both informative and readable phrases from documents. The method includes separating documents into sentences to extract sentence-level phrases and separating the documents into segments of a fixed length to extract segment-level phrases. The method further includes using a PMI and entropy model to also extract readable phrases from the documents. The method includes comparing the sentence-level phrases, the segment-level phrases, and the readable phrases to generate key phrases for the documents.

The invention discloses a sign language translation system and method based on lightweight mask enhancement, and the method comprises the following steps: 1, employing an inter-frame difference technology to recognize and reject redundant frames having no substantial contribution to motion expression, and further distinguishing the foreground and background information of sign language motion through a local dynamic mask module; step 2, inputting the sign language video optimized by the lightweight mask module to a visual module; accurate and comprehensive visual features are generated; 3, converting the visual features obtained from the visual module into language features suitable for language processing; and step 4, the language module receives the language features converted by the mapping module, and processes the language features in combination with an encoder-decoder model mBART to improve the quality of the output text. According to the method, the accuracy and efficiency of sign language-to-spoken language translation are remarkably improved; the method not only optimizes the use of computing resources, but also greatly enhances the readability and semantic coherence of translation results, and provides powerful technical support for realizing more popular and convenient sign language communication.

The invention provides a large model code translation method and device fusing code functions and styles, and relates to the technical field of natural language processing. The method comprises the steps that code pairs composed of source codes and target codes are obtained from an online programming platform, the code pairs are processed according to similarity retrieval, fine granularity scoring and difference testing, and a function consistency data set is constructed; performing functional learning training on the large model according to the functional consistency data set and an instruction fine tuning method to obtain a large model subjected to functional learning training; obtaining a source code, generating positive sample translation and negative sample translation of the source code, and constructing a style-oriented data set; and according to the style-oriented data set, style learning training is carried out on the large model subjected to function learning training, and a trained code translation large model is obtained. According to the method, a low-cost and high-efficiency code translation model is developed around a large-scale language model, and the correctness and readability of translated codes are enhanced.

The invention belongs to the technical field of information retrieval, and particularly relates to a library document abstract generation method based on deep learning model construction, which comprises the steps of data preprocessing, model construction, semantic understanding and core viewpoint extraction, abstract generation and optimization and quality evaluation. According to the library literature abstract generation method based on deep learning model construction, the mixed labeling technology is combined with rules and deep learning, the advantages of the rules and the deep learning are fully played, the rules provide a basic framework, the deep learning makes up for the deficiencies of the rules, and domain term meanings can be more accurately understood; the domain entity knowledge base ensures accurate recognition and relation understanding of entities in literatures through a rigorous extraction and disambiguation method, a foundation is laid for high-quality abstract generation from the semantic level, and in addition, through accurate understanding of domain terms and effective capture of long text semantic association, the abstract generation efficiency is improved. The accuracy, the integrity and the readability of the generated abstract are improved, so that the requirement of a user for quickly acquiring the key information of the literature is better met.

A pre-training language model-based summarization generation method, including: acquiring text information of a summarization to be generated, and performing multi-feature weighted language pre-training processing on the text information to obtain a candidate summarization; inputting the candidate summarization into a pre-training language model to obtain output data of the pre-training language model; and inputting the output data of the pre-training language model into a decoder model to obtain a target summarization. The pre-training processing is performed on the text information of the summarization to be generated, and then the target summarization can be generated after processing performed by an encoder and a decoder, improving the quality and readability of the summarization.

The invention discloses an optimization method and device for a natural language processing model, and relates to the technical field of information security, and the method comprises the steps: reducing the text confusion degree through employing an iterative beam search mode, and guaranteeing the text quality at the same time, i.e., carrying out the quantitative evaluation and screening of readability and usefulness in a processing process, and improving the processing efficiency. And performing model fine tuning by using the optimized and transformed mixed data to reduce the influence of harmful texts in the third-party data on the model security alignment, thereby improving the security performance and the alignment degree of the large model, and solving the problem that the security alignment of the large language model cannot be effectively coped with in the related technology, and the security alignment efficiency is improved. The technical problem that possible harmful information damage in the third-party data is difficult to fundamentally reduce is solved, and the technical effects that the safety performance and the alignment degree of the large language model are improved, it is ensured that the large language model can give safe and reliable response when facing the data safety problem, and harmful information is prevented from being provided are achieved.

The invention discloses a code static review method and system based on a fine tuning large language model, and belongs to the technical field of software static testing. The method comprises the steps that a code static rule set is collected and preprocessed, standard rules are screened according to programming languages and project requirements, and grading classification and duplicate removal are carried out; the method comprises the following steps: constructing a positive and negative example program data set in three modes of open source code analysis, fault implantation and large language model generation, and integrating into structured training data in combination with a template; selecting a base model matched with a code reasoning task, optimizing model parameters by adopting a LoRA fine tuning method, and predicting code rule violation conditions through output layer design; and evaluating model performance based on the indexes. According to the method, the efficiency of static review is enhanced, including readability, friendly interaction and expandability of reports, false reports and missing reports are reduced by improving the ability of code context understanding, and cross-language and cross-platform portability is realized.

The present invention relates to the technical field of data security, and in particular to an adaptive masking method and system, and a device and a medium. The method comprises: firstly, acquiring a keyword of a current file to be masked of a sender user; then, on the basis of the keyword, performing addition, deletion and modification on a current sensitive word library to obtain a new sensitive word library; and finally, generating a regular expression on the basis of the new sensitive word library, determining the positions of sensitive words on the basis of the regular expression, so as to obtain the sensitive words, and performing masking to obtain a masked file. Therefore, masking of a plurality of types of data is realized, and data types before and after masking remain unchanged, thereby ensuring the readability of a masked file while guaranteeing the security of masked data, and further increasing the masking speed; on the basis of different roles of receiver and sender users, the masking intensity is adaptively selected, thereby preventing important information from being leaked to untrusted persons, and overcoming the defect of an existing masking algorithm failing to resist collusion attacks and brute-force enumeration attacks; and multiple threads are used to concurrently process sensitive word retrieval and sensitive-word masking operations, thereby greatly increasing the masking speed.

The invention discloses a test case editing scheme oriented to the automatic test industry, and the scheme realizes efficient editing, analysis and execution of test cases through layered architecture design. The scheme comprises the following steps: (1) dividing a test case into a measurement abstraction layer and a hardware abstraction layer; (2) in the measurement abstraction layer, constructing a test index sequence and a corresponding index operation sequence to form a test process file; (3) in the hardware abstraction layer, establishing a mapping relation between the operation keyword and the underlying instrument driver, and forming an instrument configuration file; and (4) the test process file and the instrument configuration file are converted into objectified description, so that flexible editing and automatic script conversion of the test case are realized. The method has the beneficial effects that (1) the readability and editing efficiency of the test case are improved, and the workload of manual script conversion is reduced; (2) the generalization ability of the test scheme is enhanced, and the method is suitable for various test devices and test scenes; and (3) flexible configuration and efficient reuse of test resources are realized through a layered architecture and a standardized file format.

The invention provides an automatic wrong question typesetting method, a printer and a learning system, and relates to the technical field of education, the method combines a binding demand with a typesetting scheme, so that the obtained target typesetting scheme can adapt to subsequent binding operation, the problem that a user reads difficultly due to the fact that a question is shielded by binding is avoided, and binding robustness is enhanced. Moreover, according to the method, the target typesetting scheme is jointly generated in combination with the content features of all the questions in the printing task, the printing paper features, the analysis result and the binding demand information, typesetting automation can be achieved, a user does not need to arrange typesetting manually, human errors are reduced, typesetting efficiency and the content adaptability of typesetting pages are improved, and the typesetting efficiency is improved. The typesetting quality is improved, cognitive errors caused by typesetting disorder of the user are reduced, and the readability is enhanced. Moreover, the paper utilization rate can be increased, and resource waste is avoided. In addition, the analysis result is introduced, so that the target typesetting scheme of the printing task is more comprehensive and more targeted, and the user experience is improved.

The embodiment of the invention discloses a teaching note generation method and device based on a large language model and a medium, belongs to the technical field of artificial intelligence, and solves the problems that classroom notes generated in the prior art are lack of content structuring and knowledge relevance, poor in readability and difficult to be used for effective review. The method comprises the following steps: acquiring a teacher voice signal in a classroom teaching environment in real time through pickup equipment, and converting the teacher voice signal into an initial text stream through a preset voice recognition model; taking the initial text flow as query input, and performing semantic similarity retrieval on the initial text flow and a preset course knowledge base to obtain a plurality of related knowledge fragments; generating an enhanced prompt based on the initial text stream and the plurality of related knowledge fragments, and generating a structured classroom note based on the enhanced prompt and a preset large language model; and outputting the structured classroom note to a user interface, and optimizing a classroom teaching note generation process based on feedback information and modification information of the user.

The present application relates to the field of text detection technology, and in particular to a text detection method and system. The method comprises: obtaining text usage requirements; obtaining rich text historical archives based on the text usage requirements; parsing the rich text historical archives to determine structured multi-type rich texts; determining detection requirement-oriented information based on the text usage requirements and the structured multi-type rich texts; judging whether there is multimedia fusion content in the detection requirement-oriented information; if so, determining the detection result information set based on the multimedia fusion content; and outputting the detection result information set as a structured text. This ensures the accuracy and reliability of data and reduces the interference of irrelevant data; makes information more convenient for data analysis and processing, improves the practical value of information, makes the detection result information set readable and understandable, facilitates users to quickly obtain information, facilitates computer processing and analysis, and provides support for subsequent data mining and application.

This disclosure proposes a dialogue processing method, a training method for a question rewriting model, and a device thereof. The dialogue processing method includes: when information is missing in the user question, inputting the user question and historical dialogue content into the encoding layer of the question rewriting model to obtain the semantic representation vectors of each word in the historical dialogue content and the semantic representation vectors of each word in the user question, as well as the keyword scores of each word in the historical dialogue content; then, the decoding layer accurately selects keywords semantically related to the user question from the historical dialogue content based on the semantic representation vectors of each word in the user question, the semantic representation vectors of each word in the historical dialogue content, and the keyword scores, and explicitly generates a fully rewritten user question based on these keywords. This accurately supplements the missing information in the user question, achieving completeness and readability of the user question content and improving the user experience.

The embodiment of the invention provides a note generation method and device and computer equipment. The note generation method comprises the following steps: performing voice recognition on data of a guidance process to obtain text data; extracting a guidance summary according to the text data; retrieving extension information associated with the guidance summary; and generating a note of the guidance process according to the expansion information and the guidance summary. According to the embodiment of the invention, the readability and practicability of the generated notes can be improved.

The application discloses a structured legal document generation method and system based on a large model and a knowledge base enhancement, and belongs to the field of structured legal document generation. The method comprises the following steps: acquiring a case description and a document type according to user input; retrieving a legal regulation knowledge base based on the case description to obtain a relevant law article set related to the case description; retrieving a legal document template knowledge base based on the document type to obtain a legal document template set related to the document type; in the case that there is a legal document template suitable for the case description and consistent with the document type in the legal document template set, generating a legal document of the user input based on the user input, the relevant law article set and the legal document template; otherwise, performing network search according to the document type to generate a structured legal document. The application can improve the readability of the legal document and provide a more intuitive reading experience for the user.

The invention discloses a few-sample readability evaluation method and system, electronic equipment and a storage medium, and belongs to the technical field of natural language processing. The method comprises the steps of obtaining a to-be-evaluated text, and constructing a global prompt with an independent structure and at least one local prompt for the to-be-evaluated text; combining the text with each prompt, inputting the combined text and prompt into a pre-training language model, and extracting global and local prompt feature representations; respectively calculating similarity distribution between each prompt feature representation and a plurality of preset static category prototypes, wherein the static category prototypes are kept fixed in model training; and performing joint modeling on the similarity distribution from different prompts based on a Bayesian probability fusion mechanism, generating fused posterior probability distribution, and determining the readability level of the text according to the fused posterior probability distribution. According to the method, multi-dimensional language features can be effectively modeled, tag semantic fuzziness is properly processed, and high-precision and high-robustness readability evaluation is kept in a few-sample scene.

The invention relates to an official content credibility verification-oriented conditional trigger type text watermarking method, which comprises the following steps of: extracting confidence of a large language model, embedding a hierarchical trigger watermark, dividing a pseudo-random red and green set, detecting statistical characteristics and judging thresholding authenticity. Implicit watermark embedding of an official text and rapid discrimination of a public opinion end are realized. According to the method, the natural readability of the official text and the detectability of the watermark are both considered, the detection Z-score can reach 7.49, the green set hit proportion reaches 69.21%, and the method is suitable for the scenes of government affair document release, news draft clarification, network public opinion management and the like, and has good expansibility and engineering practicability.

The invention discloses a compact family structure text description grammar with high reduction degree. The compact family structure text description grammar comprises a family expression rule and an individual information expression rule. In the family expression rule, parents and children are defined as a group of family members, the parents, the parents and the children and the children are separated by different symbols, and after a group of family data is ended, different symbols are used for expressing ending description. The individual information and the information identifier in the individual information expression rule are respectively expressed by specific characters. The invention defines a structured plain text format for recording family pedigree. The core of the method is to imitate the layout of family trees: intergenerations are vertically distributed, and individuals of the same generation are horizontally aligned. The design has both human readability and machine processability, on one hand, data presentation is clear and visual, and lossless editing by using a text editor is supported; and on the other hand, compared with a traditional format, the data volume of the file is remarkably reduced, and the comfort level of user experience is improved.

The invention provides a dynamic handwriting enhancement method and device based on learning content semantic analysis and electronic equipment, and relates to the technical field of electronic note processing, and the method comprises the steps: obtaining space-time dynamic data corresponding to an original note when a user uses an electronic pen to write; performing stroke unit segmentation on the spatio-temporal dynamic data, and performing feature extraction on the segmented stroke units to obtain a feature tensor of each stroke unit; and performing semantic analysis based on the feature tensor of each stroke unit, determining a target enhancement strategy based on a semantic analysis result, and enhancing the original note based on the target enhancement strategy to obtain an enhanced note. According to the dynamic handwriting enhancement method and device based on learning content semantic analysis and the electronic equipment, by understanding the knowledge essence of the written content, sensing the real-time cognitive state of the user and applying the optimization strategy matched with the knowledge type, the handwriting readability is remarkably improved, and meanwhile the personal writing style is strictly reserved.

The invention relates to the technical field of medical knowledge maps, in particular to a method for establishing an acupuncture point knowledge base by using traditional Chinese medicine ancient books. A character sequence in an ancient book is obtained and converted into split vocabularies through vocabulary segmentation, so that acupoint vocabularies are extracted. Due to the fact that the description of the acupoint vocabularies in the ancient book is different, the corresponding operation statements are determined according to the positions of the acupoint vocabularies in the character sequence. And further analyzing the relative position and similarity relationship between the split vocabularies and the acupoint vocabularies to obtain the vocabulary distribution complexity and deeply understand the context environment of the acupoint vocabularies. And comprehensively considering the number distribution of the split vocabularies and the vocabulary distribution complexity, calculating the description variable degree value of the acupoint vocabularies, and evaluating the latitude of the information expression of the acupoint vocabularies. And finally, comparing the description variable degree values, determining the retrieval attention of the acupoint vocabularies, and marking the acupoint vocabularies to construct an accurate and easily retrieved knowledge graph, thereby enhancing the accuracy and readability of the knowledge graph.

The application provides a large model code translation method and device fusing code functions and styles, and relates to the technical field of natural language processing. The method comprises the following steps: obtaining a code pair composed of a source code and a target code from an online programming platform, processing the code pair according to similarity retrieval, fine-grained scoring and differential testing, and constructing a function consistency dataset; performing function learning training on a large model according to the function consistency dataset and an instruction fine-tuning method to obtain the large model trained through the function learning training; obtaining the source code, generating positive sample translation and negative sample translation of the source code, and constructing a style guide dataset; and performing style learning training on the large model trained through the function learning training according to the style guide dataset to obtain a trained code translation large model. The application develops a low-cost and high-efficiency code translation model around a large-scale language model, and enhances the correctness and readability of translated codes.

The invention discloses an operator report PDF exporting method and system, and belongs to the technical field of report generation and document typeset.The method comprises the following steps that a data set and a field name list of a report to be exported are received; dynamically generating a header according to the field name list, and calling a PdfPTable to create a table; for each row of data in the table, estimating the height of the row and comparing the height with the remaining available height of the current page, if the height is not enough, calling document.newPage () for paging, and reconstructing a header on a new page; and for the cells, the Phrase is used for measuring and calculating the text size, and setNoWrap (false) and the minimum height are set so as to realize column width self-adaption and automatic line feed. According to the method, the report export flexibility, attractiveness and readability are remarkably improved, and the multi-scene and multi-format customization requirements of operators are met.

The embodiment of the invention provides a handwriting generation method, electronic equipment and a computer program product, and the method comprises the steps: obtaining first handwriting data of a first handwriting input operation under the condition that the first handwriting input operation is detected; obtaining characteristic parameters of the first handwriting input operation according to the first handwriting data; the characteristic parameters are used for reflecting personalized characteristics of the first handwriting input operation; and fitting the first handwriting data based on the feature parameters to generate fitted handwriting. According to the embodiment of the invention, the first handwriting data is fitted based on the characteristic parameters capable of reflecting the personalized characteristics of the first handwriting input operation, so that the unique writing style of the first handwriting input operation can be reserved in the generated fitting handwriting, and the mechanical feeling of the fitting handwriting is avoided; the beauty and readability of the fitting handwriting are improved, and meanwhile, the personalization of the fitting handwriting is improved.

The application discloses an element text processing method and device, electronic equipment and a storage medium. The method comprises the following steps: obtaining a plurality of description information of a sample object and a sample abstract marked with an element type; extracting an element vector feature of the element type and a description vector feature of each description information; taking the element vector feature and the description vector feature as inputs of a joint abstract model to be trained, and taking the sample abstract as an output of the joint abstract model, wherein the joint abstract model comprises a first model and a second model, and a correlation degree between each description information output by the first model and the element type is an input of the second model, and then the joint abstract model is trained according to a classification loss value of the first model and a decoding loss value of the second model, so as to process commodity description information of a target object to generate a commodity abstract matched with a target element type. The text abstract generated by the method has strong readability, and corresponding text abstracts can be generated according to different element types.

The application discloses a code obfuscation method, an electronic device, a storage medium and a program product, and relates to the technical field of computers. When code obfuscation is needed, a first code is obtained, the first code comprising a plurality of first identifiers; the plurality of first identifiers are subjected to anti-decompilation unrecognizable processing based on a preset mapping table to obtain a second code, the preset mapping table comprising a plurality of corresponding relationships between identifiers and symbols, and the symbols being anti-decompilation unrecognizable symbols; a plurality of reflection code templates without actual functions are determined; and the plurality of reflection code templates are added to the second code to obtain a third code. Through the above method, the first identifiers in the first code can be subjected to anti-decompilation unrecognizable processing according to the preset mapping table, so that the anti-decompilation is erroneous, and the reflection code templates without actual functions can be added, so that the code obtained through anti-decompilation loses readability without affecting normal compilation, and the security of the code is improved.