286 results about "Text file" patented technology

The invention belongs to the technical field of RAG questioning and answering, and particularly relates to an efficient and accurate RAG questioning and answering method and device based on a knowledge graph and a storage medium. The method comprises the following steps: reading contents in a text file, dividing the contents to obtain text blocks, and obtaining an initial triple by using a large language model; performing post-processing on the initial triad to obtain an optimized triad, and constructing a knowledge graph; building an index key value pair based on the optimized triad and the text block to which the triad belongs, and storing the index key value pair in a vector database; performing mixed knowledge retrieval based on the index key value pair to obtain mixed knowledge beneficial to question answering: unstructured knowledge from the text and structured knowledge from the knowledge graph; and taking the mixed knowledge as the context of question answering, and reasoning by using a large-scale language model. According to the method, the response time is remarkably shortened, the unstructured knowledge and the structured knowledge are effectively integrated, and the accuracy of low-level retrieval and high-level retrieval is remarkably improved.

Examples of the present disclosure describe systems and methods for automating the identification of events in a text file. In examples, a computing system identifies a subset of a text file that comprises an unknown event using a set of rules. Each rule of the set of rules specifying a first pattern of characters is compared to the subset of the first text file. When the set of rules does not identify the unknown event, the subset of the text file is provided to a language model to generate a new rule with a second pattern of characters and an identifier of the new rule. The system then generates an updated set of rules by adding the new rule to the set of rules.

The invention discloses an operation planning optimization modeling enhancement method, which comprises the following steps of: based on an operation planning optimization problem text file, constructing a structured problem representation containing a dependency relationship between constraints and original semantic information; according to the structured problem representation, executing logic consistency deduction by constructing and solving forward and reverse problem pairs, and generating a forward modeling result and a reverse verification set; performing multi-dimensional verification on the forward modeling result to identify potential modeling defects, and generating a structured counter-example test set for the potential modeling defects; and integrating the test performance of the reverse verification set and the structured counter-example test set, evaluating the credibility of the forward modeling result, and generating a final optimization model and a credibility evaluation report. According to the method, the logic preciseness and credibility of the automatic generation model can be improved, and quantitative reliability evaluation is provided for the model.

The invention relates to a multi-dimensional knowledge base construction and query method based on a multi-modal large model, and belongs to the technical field of AI large model application. The method comprises the following steps of: analyzing a text file containing text, table and picture contents; selecting a pre-trained large language model and a multi-modal large model, generating text abstracts and picture abstracts by utilizing the models, and enabling storage paths of the picture abstracts to correspond to storage paths of pictures one by one; selecting a text embedding model to construct a vector database and a retriever; and constructing a retrieval chain based on the multi-modal large model. According to the method, construction from a single text knowledge base to a multi-dimensional knowledge base is achieved, a multi-dimensional information retrieval method is provided, the problem of construction of the multi-dimensional knowledge base based on the vectorization technology is solved to a certain extent, and the large model output quality based on the retrieval enhancement generation technology is remarkably improved.

The invention discloses a rich text multi-dimensional difference analysis and accurate revision method and system based on AI, and relates to the technical field of artificial intelligence and data analys.The rich text multi-dimensional difference analysis and accurate revision method comprises the steps that firstly, a rich text file is analyzed, a document structure is reconstructed, vectorization and key information labeling are conducted on semantic blocks, and structuralization and semantic enhancement of a text are achieved; the method comprises the following steps of: firstly, carrying out text surface layer, semantic deep layer and structure association dimension difference analysis in parallel, then carrying out system knowledge graph-based revision influence degree evaluation on the basis of difference analysis, tracing associated terms through penetrating analysis, calculating an influence degree score and early warning potential conflicts, and automatically generating accurate revision suggestions; visual difference preview and one-key generation of a revised draft are provided, a closed loop from problem discovery to problem solving is realized, and the efficiency and accuracy of system revision are greatly improved.

Systems, computer program products, and methods are provided for code revision impact analysis. The method includes generating a system map based on data received from a plurality of network devices; receiving a data transmission including a text file; processing the text file via a natural language processing engine, where an output of the natural language processing engine comprises a plurality of expected updates; determining, based on the system map, at least one downstream effect of the plurality of expected updates; and performing a remedial action.

The invention discloses an ANTLR-based SysMLv2.0 text model synchronization method, which comprises the following steps of: S1, constructing a syntactic analysis framework of a SysMLv2.0 language on the basis of the ANTLR, and generating a Leyer and a Parser for analysis; s2, analyzing the original SysMLv2.0 text, and generating an abstract syntax tree AST containing structure, behavior and constraint elements; s3, constructing a semantic model structure based on AST mapping; s4, respectively constructing an original semantic model and a target semantic model, and unifying an identifier, a namespace and a reference path format; s5, comparing semantic model structure differences to generate a difference list; s6, performing insertion, deletion or replacement operation on the original model text according to the difference type, and performing incremental synchronization; s7, deserializing the synchronized semantic model to generate a standard SysMLv2.0 modeling text; and S8, outputting the text file for loading and visual operation of the modeling tool. According to the method, a unified analysis and synchronization system is constructed, and accurate comparison and incremental updating of the SysMLv2.0 text model are realized.

Systems and methods for distributed ledger-based check verification are disclosed. In one embodiment, a method may include a bank backend computer program: (1) receiving, from a computer application executed by an electronic device, an image of a presented check as part of an electronic check deposit process; (2) performing optical character recognition on the image of the presented check; (3) generating a text file based on the optical character recognition; (4) querying a distributed ledger in a distributed ledger network to determine whether the presented check has been presented or cleared before; (5) determining that the presented check has not been presented or cleared before; (6) processing the presented check for deposit; and (7) writing the text file for the presented check to the distributed ledger.

To improve usability of a user.SOLUTION: The document information includes a document display name of a document file corresponding to a search condition, a list of detected text sentences including a search word and characters before and after the search word in a text file, most frequent image data corresponding to a page in which the search word is detected most frequently, and a jump button that can be linked to a document address.SELECTED DRAWING: Figure 10

Systems, computer program products, and methods are provided for automated detection of source code discrepancies. The method includes receiving a data transmission including a text file and a source code file; processing the source code file via a machine learning engine, where an output of the machine learning engine includes a plurality of identified updates; processing the text file via a natural language processing engine, where an output of the natural language processing engine includes a plurality of expected updates; identifying a difference between the plurality of identified updates and the plurality of expected updates; and performing a remedial action.

To improve usability of a user.SOLUTION: When a search condition including at least one search word is supplied from a user terminal, searching a local government document database for a document disclosed by a local government, searching for the search word in the text file corresponding to a document ID matching the search condition, and generating result image data for displaying a list of document information in which a document display name including the search word, a detected text sentence including the search word and characters before and after the search word in the text file, and a document link button including a link to the document file are displayed in one frame.SELECTED DRAWING: Figure 1

The invention provides a relay protection device interface language switching method and system based on a hot loading mechanism, and the method comprises the steps: constructing a variable-length language storage structural body under a C language according to the variable-length text characteristics of multiple languages in a language text file, so as to generate an instantiated object, and further binding an interface element of a relay protection device; under the condition that an operation signal of switching languages of the user interface is detected, obtaining a user target language according to the interface element index so as to read a corresponding language text file; pointing a first-level pointer corresponding to the instantiated object to the read language text file, and pointing a second-level pointer corresponding to the instantiated object to the first-level pointer so as to bind the interface element and the target language text file; analyzing a second-level pointer corresponding to the instantiated object, and mapping the second-level pointer to a first-level pointer corresponding to the instantiated object to obtain target language text information pointed by the first-level pointer. According to the method, interface language switching can be completed without restarting equipment.

This application discloses a method and an apparatus for training a model for text retrieval, a computer device, and a storage medium. Specifically, the method includes: parsing a text file and establishing a document tree; constructing a training sample according to the document tree, the training sample including a text pair and a relevance label related to the text pair; and training a model by using text and the relevance label in the training sample. In above way, valid positive and negative training samples can be established. The model can be configured for text retrieval after being trained by the text and the relevance label in the training sample, and a retrieval result is accurate and effective accordingly.

Certain aspects of the disclosure provide a method for extracting attributes from documents with varying formats, layouts and complexities. The method displays a user interface (UI) that enables a user to obtain an unstructured document from a knowledge base. The method converts the unstructured document into a text document using a text recognition. The method obtains, as output from a large language model (LLM), an extracted page attribute from the text document. The extracted page attribute contains a first type of information recorded in text on a single page of the text document. The extracted document attribute contains a second type of information recorded in text on more than one page of the text document. The method obtains, as output from the LLM, an extracted document attribute from the text document. The extracted page attribute and the extracted document attribute are displayed in the UI.

The invention provides a sequence decision strategy acquisition method and system for high-neuroplasm crowds. The method comprises the following steps: establishing a full-screen drawing canvas, and loading a task set comprising a decision map type, an emotion titer ratio, a reward parameter, a random seed and a sequence control parameter; collecting user gender selection and binding a corresponding emotion picture library; generating a decision map based on a Markov decision principle under the control of the parameter set, defining nodes and connecting edges, randomly scheduling and laying emotion pictures according to a titer ratio, and assigning rewards to form an emotion and reward coupling environment; based on the environment, recording a mouse position and a mouse path in an execution exploration stage, recording key pressing time, an action direction and an award in an execution stage, and establishing step sequence association according to a timestamp to form sequence decision data; a Latin square balance control task sequence is adopted, and the process is propelled under the trigger of a user; and finally, filing the data in a binary system and a text file for subsequent analysis. According to the invention, the authenticity and repeatability of data are improved.

A hypergraph user interface for interaction with a system for integrative causal modeling from natural language text and including a natural language text input element configured to receive a designation of a natural language text file or such a file. The input element causes the designation or natural language text file to be communicated to at least one natural language processing module. The hypergraph user interface further includes a causal hypergraph visualization element representative of directed hypergraph data generated from the at least one natural language processing module. Selection of a hypergraph element of the causal hypergraph representation causes the hypergraph user interface to communicate an indication of the hypergraph element to a querying module and generate an informative element indicative of at least one of a causal link, a causal statement, or a sentence of the at least one natural text file on which the hypergraph element is based.

The invention relates to the technical field of databases, in particular to a MySQL binary log incremental recovery import failure position quick positioning method, which comprises the following steps of: 1, executing a flush-logs command in a target database server to generate a target end latest binary log file, in the conversion process, the line boundary of the event and the relative sequence of endlogpos and CRC32 checksum are kept unchanged; step 2, scanning the binary log text file of the target end line by line by adopting a co-occurrence detector, and determining an endlogpos value analyzed in the co-occurrence of the source end in the stable pairing as a final recovery breakpoint position; and step 3, the controller generates an increment recovery parameter according to the final recovery breakpoint position, and if the import is successful, the recovery breakpoint position and the common line number mapping of the source end and the target end are recorded. According to the method, the automation degree and accuracy of incremental recovery are remarkably improved, and data loss or repeated import caused by inaccurate breakpoint positioning is reduced.

A method for creating an advertisement based on user input includes displaying, by a public display device, a link for inputting user-input text comments about a predetermined merchandise; in response to receipt of a user-input text comment, performing, by a server, a filtering operation so as to obtain a filtered input string; feeding the filtered input string into a generative neural network model, so as to obtain an advertising text file; generating a user-related advertisement for the predetermined merchandise based on at least the advertising text file, and transmitting the user-related advertisement to the public display device for displaying.

Systems and methods are disclosed herein for generating updated descriptions of items based on analyzing candidate embeddings of semantic representations of item descriptions. The system may obtain a text file describing an item. The system may provide the text file to a generative language model to generate semantic representations of the text file. The system may generate, based on the text file, candidate embeddings in an embedding space. The system may obtain embeddings associated with existing items. The system may determine subsets of the embeddings within a threshold distance. The system may compare the subsets. The system may determine attributes associated with a candidate embedding based on the comparison. The system may generate an updated text file based on the attributes.

Various embodiments of the present disclosure provide a text interpretation technique. The text interpretation technique includes generating a plurality of text segment embeddings from a plurality of input text documents and identifying prompt embeddings associated with a predictive task. The technique includes generating task-specific similarity scores for the plurality of text segments based on a comparison between the plurality of text segment embeddings and the prompt embeddings, The technique includes identifying a set of task-specific text segments from the plurality of text segments based on the plurality of task-specific similarity scores and training a target machine learning model based on the set of task-specific text segments.

The invention relates to a method and a system for creating an advertisement based on user input. The method for creating the advertisement based on the user input comprises the following steps: displaying a link for inputting a user input text comment about a predetermined commodity on a public display device; in response to receiving a user input text comment, the server performs a filtering operation to obtain a filtered input character string; feeding the filtered input string to a generative neural network model, thereby obtaining an advertisement text file; generating a user-related advertisement for the predetermined commodity at least based on the advertisement text file; and sending the user-related advertisement to a public display device for display. In this manner, the resulting user-related advertisement may be more readily accessible to the public because the content is obtained directly from the user's user experience.

The invention discloses an XML-based video template language definition method and device, and the method comprises the steps: creating a VTL video template, and setting a plurality of XML attributes for a template element for definition; embedding one or more scene elements in the template element through a parent-child relationship of XML (Extensible Markup Language); various content elements are nested in each scene element; and introducing a plurality of logic elements with the same level as the template element, wrapping the template element and the plurality of logic elements by using a root element, and realizing dynamic logic processing and data binding in cooperation with a dynamic calculation expression so as to generate the VTL video template capable of being analyzed and rendered. Through the method provided by the invention, the VTL completely, accurately and flexibly defines a complex video item as a standardized text file by utilizing the structuralization, extensibility and self-description characteristics of the XML, so that comprehensive description from static layout to dynamic content generation is realized; standardized description of the video content is achieved, high dynamic performance is achieved, secondary editing and cross-terminal unified construction are easy, and the efficiency and flexibility of video production are improved.

A computer system includes circuitry for executing operations, including receiving an input text file comprising metadata, receiving risk enrichment data, and generating an affinitized data set based on the input text file and the risk enrichment data. Generating the affinitized data set includes the operations of determining a context-indicative keyword, determining a synonym of the context-indicative keyword, searching the input text file for the context-indicative keyword or the synonym, identifying the context-indicative keyword or the synonym in the input text file, and generating the affinitized data set, the affinitized data set comprising a risk descriptor determined based on the at least one of the plurality of context-indicative keywords and the metadata.

The embodiment of the application discloses a data conversion method, system and device, the method comprises the following steps: when a trigger condition corresponding to a scheduling task is met, triggering the scheduling task, and transmitting task parameters to a push task main program; through the push task main program, target data in a database is queried based on the task parameters in a stream mode, the target data is parsed according to a template file, and the target data is written into a landing shard data file; after the target data in the database is queried and written into the landing shard data file, the total number of the target data under each data mode is determined, a text file corresponding to each data mode is generated according to the total number of the target data under each data mode and each landing shard data file under each data mode; and the text file corresponding to each data mode is compressed, and the compressed text file is uploaded to a target server. The method can ensure the timeliness and accuracy of data pushing, and realizes decoupling of the database and a data consumer.

The invention provides a rock mass fracture network medium reconstruction method and system based on image tracking, and belongs to the technical field of fracture network reconstruction, and the method comprises the steps: obtaining a to-be-reconstructed fracture network, and carrying out the preprocessing of the to-be-reconstructed fracture network, and obtaining a binary image; skeletonizing the binary image to obtain a single-pixel wide fissure main ridge; identifying end points and intersections of the fracture skeleton based on a neighborhood connection relation, and performing constrained graph tracking by taking the end points and the intersections as graph nodes, taking a connection skeleton sequence between the end points and the intersections as graph edges and taking the end points as starting points to obtain an ordered vertex coordinate sequence of each fracture broken line; and performing multi-point coordinate calibration on the ordered vertex coordinate sequence, mapping the ordered vertex coordinate sequence to a target two-dimensional coordinate system, generating a polyline command text file based on the mapped coordinates, and reconstructing the fracture network medium by using the polyline command text file. The method is oriented to data flow of heterogeneous software, a special interface is not needed, batch processing is supported, and the efficiency and accuracy of cross-platform modeling are improved.

In various embodiments, systems and methods for ledger-based cookie management are provided. Rather than store cookie data as text files on the local device drive, cookie data is recorded to a blockchain technology cookie ledger store on a network resource. When a client application (e.g., a browser) on the UE is directed to a cloud-based service, and that cloud-based service calls for access to a cookie, that call is processed by a cookie gateway executing on the UE. The cookie gateway may verify authenticity of the user and generate a cookie access token that it transmits to the cloud-based service. The cloud-based service may use the cookie access token to locate the cookie ledger and access one or more records storing cookie data used by the cloud-based service. The cookie access token may expire upon termination of the session between the user equipment and the cloud-based service.

The invention relates to the field of marine carbonate rock sequence research, and discloses a marine carbonate rock reservoir high-frequency sequence division method and system.The method comprises the steps that firstly, a GR curve is exported into a text format, and a GR curve text file is obtained; the method comprises the following steps: importing an Acycle, and carrying out pretreatment; performing spectral analysis, screening a curve with a standard Milankovitch cycle ratio, and calculating the average number of each level sequence developed by the layer section; carrying out wavelet transformation on the GR curve by utilizing MATLAB (matrix laboratory), and separating out wavelet curves of different levels; carrying out trending processing on the GR curves of the multiple target layers through INPEEA method processing to obtain a contrast curve; carrying out wavelet signal curve level matching and layer sequence division, and carrying out contrast inspection; according to the method, effective signals reflecting different levels of cycle changes can be extracted, high-frequency cycles are accurately divided, the influence of subjectivity and multi-level sequence superposition interference on sequence division is effectively eliminated, and a theoretical basis and service are provided for deep dolomite reservoir distribution prediction.

The embodiment of the invention provides an ancient book proofreading method and device and a storage medium. In the method, each ancient book page of an ancient book file is divided into a plurality of ancient book segments, a text segment corresponding to each ancient book segment is generated, and a first ancient book character of each ancient book segment corresponds to a second ancient book character in the corresponding text segment; according to the first ancient book character in any ancient book segment, performing letter proofreading on the second ancient book character in the text segment corresponding to any ancient book segment to obtain first proofreading information; according to the syntactic structure in any ancient book page and the first proofreading information in the text page, performing word and sentence proofreading on the text page to obtain second proofreading information; and according to the first and / or second proofreading information in the text file corresponding to the ancient book file, carrying out grammar proofreading on text contents with impaired meanings in the text file to obtain a corrected text file. In this way, the ancient book recognition text can be corrected from multiple dimensions of characters, words, sentences and semantics, and the accuracy of the recognition text is improved.

The invention relates to an information synthesis system comprising: - at least one information receiving means (R1, R2) for receiving text (T), audio (A) and / or video (V) files; - a means (C) for converting a multimedia file (A, V) into text in order to obtain at least one converted file (FC); - a first artificial intelligence (AI) model for analyzing the text file (T) and / or the converted file (FC), and for generating a mental map (CM) on the basis thereof, the first artificial intelligence (AI) model being constructed by machine learning from construction data comprising reference files (T1, T2, …, Ti, A1, A2, …, Ai, V1, V2, …, Vi), as well as corresponding reference mental maps (CM1, CM2, … CMi); the invention further relates to a method and a computer program based on such a system.

The present invention discloses a method for processing images and texts of a laser control system. The method comprises the following steps: obtaining a machine workstation photograph taken by a fisheye camera, correcting the machine workstation photograph to obtain a background image, importing the image and text into a server and obtaining a file format of the image and text, performing image processing on the image and text according to the file format to obtain a design drawing, constructing a canvas according to the background image, performing a first graphic matching between the background image and the canvas to obtain a canvas background image, performing a second graphic matching between the design drawing and the canvas background image to obtain a canvas design file, determining a graphic path and graphic processing parameters of the canvas design file, classifying the canvas design file into layers based on the graphic path and the graphic processing parameters to obtain a file to be processed, adopting a fisheye camera to correct the image to improve the image processing accuracy, importing different image and text file types, performing graphic matching between the background image, the canvas and the design drawing, thereby forming multiple styles of files to be processed, and meeting different needs of users for laser engraving.