290 results about "Text matching" patented technology

The invention relates to a weak supervision instance segmentation method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining a training image set and an image-level label, and generating a pseudo-semantic segmentation label; extracting multi-scale image features in the training image set, performing prompt contrast learning based on the multi-scale image features and the pseudo-semantic segmentation labels, and generating contrast loss and a pixel-text matching graph; splicing the multi-scale image features and the pixel-text matching graph to generate enhanced features, and training an instance positioning module based on the enhanced features to generate an instance center point and positioning loss; generating a semantic mapping graph through the segmentation network, and calculating self-matching loss of prompt guidance; calculating the total loss of the model based on the obtained loss, and performing iterative training on the instance segmentation model based on the total loss of the model to generate a target instance segmentation model; and generating a target instance segmentation result based on the to-be-processed image and the category text through the target instance segmentation model. According to the invention, the precision of instance segmentation is improved.

The invention discloses a text matching positioning method, system and device and a medium. The method comprises the steps of obtaining an original long text and segmenting the original long text into ordered text blocks; selecting the first N text blocks to form a first candidate set based on the common substring sorting of the to-be-matched text and each text block; intercepting the first candidate set according to head and tail characters of the to-be-matched text to obtain a second candidate set; calculating similarity scores of the text blocks in the second candidate set and the to-be-matched text through a sequence matching similarity algorithm; combining the scores with the first candidate set to determine a matching result and output a position, and judging that no matching exists when all the scores are lower than a preset minimum threshold value. According to the method, the limitation of a traditional matching mode is broken through through primary screening of common substrings, accurate interception of head and tail features and multi-layer processing of a sequence matching similarity algorithm, complex scenes such as text missing and semantic equivalent expression can be effectively handled, the accuracy and robustness of text positioning are improved, and the user experience is improved. The method can be widely applied to the fields of document management, information retrieval, academic analysis and the like.

The present disclosure provides a training method and apparatus for an image-text matching model, a device and a storage medium. The method includes: acquiring a positive sample and a negative sample; where the positive sample includes text and an image, the text in the positive sample is used to describe content of the image in the positive sample; the negative sample includes text and an image, the text in the negative sample describes content that is inconsistent with content of the image in the negative sample; training the image-text matching model by using the acquired positive sample and the acquired negative sample based on a manner of contrastive learning; where the image-text matching model is used to predict, for an input image and input text, whether the input text is used to describe content of the input image.

The invention relates to the technical field of multi-modal cross-border advertisement generation, and discloses a multi-modal cross-border advertisement generation method and device, electronic equipment and a storage medium, and the method comprises the steps: greatly improving the image-text matching accuracy through introducing a cultural feature embedding and cross-modal alignment mechanism, and combining modeling and automatic screening, and improving the image-text matching efficiency. According to the method, high unification of advertisement contents in the semantic level is ensured, meanwhile, the defect of insufficient culture adaptability of a general template is effectively overcome, advertisement schemes violating regional culture rules can be accurately recognized and filtered by means of a culture feature vector library and adaptation index calculation, and the risk of misuse of culture is reduced to the minimum. The method has the beneficial effects that the audit passing rate and brand safety of the advertisement in a specific market are remarkably improved, full-link automation from content generation to final synthesis is realized, the generation efficiency is greatly improved, and the labor cost is reduced.

The invention discloses a similarity calculation method based on semantic editing fusion. The method comprises the following steps: calculating global semantic similarity of a source text and a target text by utilizing a semantic coding model; obtaining a first keyword list of the source text and a second keyword list of the target text through word segmentation processing, taking each first keyword in the first keyword list as a target word, and respectively forming a plurality of to-be-compared word pairs with a corresponding word in the second keyword list and a plurality of adjacent second keywords; based on the to-be-compared word pairs, calculating the semantic similarity between each group of target words and the corresponding words by utilizing a semantic coding model, and determining the editing distance between the source text and the target text; and determining the local semantic similarity of the source text and the target text according to the editing distance, and combining the global semantic similarity to obtain the final text matching similarity. According to the method, the problem that text matching only depends on character-level surface matching and neglects semantic association between word pairs is solved, and the text matching precision is improved.

The invention discloses a text matching method and device based on probability distribution, equipment and a storage medium, and relates to the technical field of natural language processing, and the method comprises the steps: obtaining semantic feature distribution of each professional knowledge text from a professional knowledge base, obtaining a knowledge probability distribution set, abstracting the intrinsic features of the text through probability distribution, and obtaining a knowledge probability distribution set; and standardized representation of the knowledge base is realized. The text input by the user is obtained, the corresponding user text probability distribution is calculated, and the fault tolerance of text matching is remarkably improved. The similarity distance between the user text probability distribution and each distribution in the knowledge probability distribution set is calculated, and the overall semantic similarity instead of local matching is measured through the distribution distance, so that the comparison process can tolerate the distribution offset. Finally, a text matching result is determined according to the minimum value of the similarity distance, the effect of stably retrieving semantic related professional knowledge under high noise is achieved, and fault tolerance and matching precision are remarkably improved.

The invention provides a calculation amount drawing-oriented large model training data labeling method, device and equipment, which are used for labeling key information of a construction engineering cost calculation amount drawing. The method comprises the following steps: displaying a to-be-labeled calculation amount drawing, determining a to-be-identified area, and identifying to obtain text content and position information of a text block; displaying a label list according to the keyword field; after a user selects the tag, activating the interaction state of the to-be-identified area, so that the text block is presented as an operable hot area; after the user selects the hot area, generating a key value pair by using the selected label and the hot area text; and collecting the key value pairs of the to-be-identified areas to generate answer content, and filling the answer content into an answer part of a preset question-answer template to obtain a labeling result. By adopting the method, the field and text matching can be quickly completed, the structured annotation data is generated, the annotation efficiency and consistency are improved, and the method can be conveniently and directly used for large model training.

This application discloses a text matching method and related equipment. The related equipment may include a text matching device, an electronic device, a computer program product, and a computer-readable storage medium. This application involves acquiring at least one text pair to be matched, concatenating the texts to be matched in the text pair to obtain concatenated text, extracting semantic features from the concatenated text to obtain text word features for each text word in the concatenated text and text features of the concatenated text, fusing each text word feature with the text features to obtain fused text features, and then using a convolution kernel of at least one size to extract multi-dimensional features from the fused text features to obtain target convolution features. Based on the target convolution features and the fused text features, the matching result of the text pair to be matched is determined. This scheme can improve the accuracy and efficiency of text matching.

This application discloses an image-text matching method, device, and storage medium. The image-text matching method includes: extracting global visual features using the visual coding features of the target image, and extracting global text features using the text coding features of the target text; using the text coding features, generating importance parameters corresponding to each visual feature channel in the global visual features; removing features corresponding to visual feature channels whose importance parameters do not meet the requirements from the global visual features to obtain enhanced visual features; and performing feature matching on the enhanced visual features and the global text features to obtain a matching result for the target image and the target text. This method can improve the accuracy of image-text matching.

This invention provides a method, system, device, and storage medium for generating advertising images in OTA (Online Travel Agency) scenarios. The method includes: receiving an image generation request containing merchant attributes and advertising copy; parsing the advertising copy using a large language model, extracting core keywords, and breaking it down into at least one image generation task based on merchant attributes; calling a finely tuned image generation model to perform image generation operations based on the image generation tasks; performing triple standardization constraint checks on the generated images, including size, content, and compliance; uploading qualified images to the platform server and providing feedback on the results. This invention enables fully automated generation from copy to advertising images, significantly improving image-text matching accuracy and image generation efficiency, reducing labor costs and operational risks, and meeting the high-frequency, personalized batch advertising material production needs of OTA merchants.

This invention discloses a long text matching method combining noise filtering and a divide-and-conquer strategy. The method includes: constructing a long text matching model, which comprises a keyword extraction layer, an association extraction layer, and a filtering layer. The keyword extraction layer extracts keywords from the text; the filtering layer filters noise from the text based on sentence similarity to obtain a denoised text sequence; and the association extraction layer further removes keywords from the denoised text sequence to obtain the remaining associated text. The long text matching model is trained with the optimization objective of minimizing a set overall loss function, which reflects the global matching distribution and combines the keyword and association matching distributions. For the target text, real-time text matching is performed using the trained long text matching model. This invention improves the generalization ability and accuracy of text matching.

This application provides a data processing method, apparatus, device, and medium. The method includes: acquiring a first text containing business text data; performing a risk assessment on the first text to obtain a risk category result corresponding to the first text; if the risk category result is a first risk category, acquiring keywords of the business text data and searching for the keywords in a standard database; if a request text matching the keywords is found in the standard database, determining the feedback text corresponding to the request text as the business processing result corresponding to the business text data; if no request text matching the keywords is found in the standard database, performing text search processing on the business text data in a target knowledge graph to obtain a business processing result matching the business text data. Implementing this application embodiment can improve the security of text data.

The application relates to the technical field of natural language, and discloses a text error correction processing method, which comprises the following steps: obtaining a text to be corrected; performing feature extraction on the text to be corrected through a feature layer of a text error correction model to obtain position features, image features and font features; performing error labeling on text characters in the text to be corrected according to the position features, the image features and the font features through a labeling layer of the text error correction model to obtain error characters; performing stroke decomposition on each error character to obtain character strokes; performing character matching on the character strokes of each error character through a preset homophone dictionary to obtain candidate characters; screening target characters from all the candidate characters, and replacing the error characters with the target characters to obtain a target text. Through the character matching on the character strokes of each error character, the character matching using stroke information is realized, and the accuracy of text error correction in the information input process in the insurance field is improved.

The invention provides a text matching method and device, equipment and a medium. Relates to the technical field of data analysis, and the method comprises the steps: obtaining a to-be-matched text, cutting the to-be-matched text according to a preset root in a preset root library, obtaining a target root combination of the to-be-matched text, and aiming at each root in the target root combination, according to the text index information of the roots in a preset root index library, obtaining the to-be-matched text. And finally, according to the text index information of each word root in the target word root combination, obtaining a matching result of the to-be-matched text from the alternative text library. Through the embodiment of the invention, the obtained matching result is the preset text with more roots in the target root combination in the preset text library, so that the text matching efficiency is improved.

The application discloses a text fuzzy matching method and device; the application can acquire a word to be fuzzy matched; a target word segmentation is determined from a preset word set based on the word to be fuzzy matched, a word prefix of the target word segmentation contains the word to be fuzzy matched, and a word prefix of a first adjacent word segmentation of the target word segmentation does not contain the word to be fuzzy matched; a target document identifier corresponding to the target word segmentation is acquired based on the target word segmentation and a mapping relationship pair; a document corresponding to the target document identifier contains the target word segmentation; the target document identifier is added to a fuzzy matching set of the word to be fuzzy matched; the fuzzy matching set includes document identifiers matched by the word to be fuzzy matched; the fuzzy matching set is updated based on a second adjacent word segmentation of the target word segmentation; and a fuzzy matching result of the word to be fuzzy matched is acquired; the application can improve the retrieval efficiency by improving the fuzzy matching algorithm.

The invention provides an investment project duplicate checking method and system, a computer and a storage medium, and the method comprises the following steps: collecting declaration data of an investment project, and generating a standardized data set; semantic analysis is carried out on the standardized data set to extract semantic feature vectors, a knowledge graph related to the declaration data is constructed, associated feature vectors are generated, and fusion feature vectors are obtained through combination; according to a mean value of the fusion feature vector and the historical feature vector, calculating a semantic deviation compensation amount based on a compensation model; and calculating to obtain a target feature vector, and calculating the similarity between the target feature vector and the historical item feature vector to screen the repeated declaration data with the similarity greater than a preset threshold. According to the method, semantic and knowledge graph association are integrated through multi-modal feature extraction, the defect that semantic understanding of text matching is insufficient is overcome, feature vectors are adjusted through dynamic semantic compensation, efficient decision is finally achieved through the similarity matching step, and therefore the duplicate checking accuracy and efficiency are overall improved.

The invention relates to the technical field of ship equipment, discloses an off-line large-model intelligent bidirectional translation system oriented to navigation communication guarantee, and aims to solve the problem that navigation communication among different languages depends on networking requirements of translation software or depends on subjective judgment of accompanying translators in the prior art. According to the technical scheme, the communication content is extracted and translated by using the pre-established model, the subjectivity of manual translation is avoided, networking is not needed, and communication among different languages is realized. According to the invention, the effective voice is extracted, so that the influence of environmental noise is reduced; the voice is converted into the text, communication content can be visually displayed, and information omission is avoided. And the translation unit loaded with the text matching model integrated with the special term of the marine communication can translate special words and sentences in the marine communication, so that the translation accuracy is improved. And the audio recording unit and the storage unit can store communication contents and translation contents, so that subsequent redisk is facilitated.

This invention relates to artificial intelligence technology and discloses an image-text matching method, comprising: extracting image features from the image to be matched to obtain an image feature vector; performing weighted transformations on the text feature vector and the image feature vector based on a self-attention mechanism to obtain a first weighted result and a second weighted result; fusing the first weighted result and the second weighted result with cross-attention to obtain a fused text feature vector and a fused image feature vector; performing vector concatenation processing on the fused text feature vector and the fused image feature vector to obtain a matching feature vector; and calculating the matching probability based on the matching feature vector using a preset classification function to obtain a matching result. This invention also relates to blockchain technology, wherein the matching feature vector can be stored in a blockchain node. This invention also proposes an image-text matching device, apparatus, and medium. This invention can improve the accuracy of image-text matching.

The invention discloses an enterprise patent demand text matching method and system based on multi-modal data fusion, and relates to the technical field of data processing. The method comprises the following steps: storing multi-modal operation data in an enterprise database; based on the operation data, obtaining a plurality of first vocabulary three-primitive groups; based on the patent demand text, obtaining a plurality of demand vocabularies and a plurality of second vocabulary three-original groups; obtaining a demand intensity value corresponding to each demand vocabulary based on each first vocabulary three-original group and each second vocabulary three-original group; and based on each demand intensity value, obtaining a matching range of the patent demand text. According to the method, the matching range of the patent demand is divided based on the demand intensity value by calculating the demand intensity value of the demand vocabulary, and the final matching range is dynamically adjusted in combination with the topological distance and the intensity difference value, so that the matching result better fits the real demand of the user, and the matching efficiency and accuracy of the patent demand and the enterprise operation data are improved.

The invention discloses a voice quality evaluation method and device, equipment and a storage medium, and relates to the technical field of computers. The method comprises the steps of obtaining to-be-evaluated voice generated through text-to-voice conversion and style information for the to-be-evaluated voice, and extracting basic features of the to-be-evaluated voice; extracting text-to-speech features from the basic features; the text-to-speech features comprise any one or more of rhythm features, tone features and text matching features; calculating a style matching degree between an actual style embedding vector corresponding to the text-to-speech feature and a style embedding vector template corresponding to the style information; and predicting a speech degradation score based on the text-to-speech features, and determining speech quality based on the speech degradation score and the style matching degree. By extracting rhythm features, timbre features and text matching features, identifying specific quality defects in text-to-speech conversion; and in combination with the style matching degree, the problem of confusion of style differences and quality defects is solved, and the accuracy of voice quality scoring is improved.

This application provides a text processing method, apparatus, electronic device, computer-readable storage medium, and computer program product based on artificial intelligence; relating to fields such as cloud technology, artificial intelligence, and intelligent transportation. The method includes: performing multi-scale encoding processing on a first word vector sequence corresponding to a first text to obtain multi-scale semantic features of the first text; performing multi-scale encoding processing on a second word vector sequence corresponding to a second text to obtain multi-scale semantic features of the second text; constructing a similar semantic matrix based on the multi-scale semantic features of the first text and the second text; performing feature extraction processing on the similar semantic matrix to obtain similar semantic features; and determining the matching degree between the first text and the second text based on the similar semantic features. This application can accurately express the semantics of text, thereby effectively improving the accuracy of text matching.

A low-altitude unmanned aerial vehicle zero-sample visual language navigation method for urban scenes. Under the condition that the absolute position of the navigation target is unknown and GNSS signal is denied, the forward-looking image, depth image and overhead view of the unmanned aerial vehicle are obtained; based on the weighted recursive of the continuous historical track points before the signal loss, the relative direction of the target is calculated from the historical track and the heading correction information is generated; the forward-looking image is divided into multiple image regions, the semantic similarity of each region and the natural language task target is calculated by using a picture-text matching model, and the highest score region is selected to generate a physical position prompt; the multi-view image, task target, heading correction information and position prompt are jointly input into a multi-modal large model, the normalized decision instruction containing action type and amplitude parameter is constrained by the pre-defined action set, and the autonomous navigation closed loop of visual perception, language reasoning and flight control is formed after analysis and execution. The invention realizes zero-sample open vocabulary navigation without preset coordinates and special detectors under GNSS denial.

The invention particularly relates to a network flow protocol identification method and device, and the method comprises the steps: a data package capturing module obtains a network data package from a network card in real time, and stores the data package in a cache region; the feature extraction module reads the data packet from the cache region and extracts features including source / destination IP addresses, source / destination ports, protocol types, transmission directions, retransmission or fragmentation, URLs, digital certificates, communication modes and session statistical information; the protocol identification module analyzes the extracted feature information, and performs protocol type judgment on the data packet by comprehensively using protocol specific field detection, multi-mode text matching, IP address / port mapping and a machine learning algorithm; the protocol library updating module is used for regularly updating a protocol library according to the newest evolution condition of the internet and a network protocol and the feedback of an application system; and the result output module outputs the identification result. The method solves the problems that an existing network flow protocol identification method is low in accuracy, insufficient in efficiency, limited in adaptability and the like.

The application provides a self-evolution presentation document generation method and system based on multi-modal and knowledge graph, wherein the self-evolution presentation document generation method comprises the following steps: constructing an enterprise-level picture meta-database containing a vector index; parsing a source document to obtain structured content and generating a presentation document outline; searching for matching candidate pictures in the enterprise-level picture meta-database based on the target page core content in the outline; querying a design decision knowledge graph to generate layout and color matching planning; assembling a presentation document according to the planning and synchronously generating design decision metadata; driving knowledge graph updating by displaying the design decision metadata and obtaining user modification feedback on the design; and the self-evolution presentation document generation system comprises functional modules for realizing the above method. The application solves the technical problems of poor picture quality, inaccurate image-text matching, rigid design, high copyright risk and difficulty in continuous learning and evolution from use in the existing automatic presentation document generation technology.

The application discloses a part surface quality detection method, device, equipment and medium, and relates to the technical field of intelligent manufacturing, which comprises the following steps: acquiring a part image containing surface information of a part to be detected, selecting a defect description text based on detection requirements, and constructing a defect detection task instruction; inputting the part image and the defect detection task instruction into a preset quality detection model to calculate a first affinity of the part image and a second affinity of the defect detection task instruction through each matching unit, selecting a target matching unit from each matching unit based on the first affinity and the second affinity; activating the target matching unit to generate an initial graphic-text matching graph; reading image edge information of the part image through the preset quality detection model to obtain diffusion modulation parameters, and iteratively processing the initial graphic-text matching graph based on the diffusion modulation parameters, and then outputting a binary segmentation image containing a defect prediction result. High-precision detection of the part is realized.

The invention provides a multi-modal large model fine-tuning corpus production method for the field of planning and natural resources, and aims at solving the problems that existing corpus lacks professional semantics, manual annotation is low in efficiency, and no standardized process exists. According to the method, an exclusive VQA corpus is generated through four core modules including construction of an industry exclusive cognitive task and VQA template library, image-text pairing generation, template-driven VQA sample generation and automatic quality control optimization in combination with an industry planning standard and a cognitive hierarchy, namely perception-reasoning-association-application. The method comprises the following steps: firstly, automatically extracting visual elements and policy texts of a planning graph, and matching a professional template to generate candidate Qamp; the method comprises the following steps: A, finally outputting a standardized corpus package through semantic detection and expert re-check optimization; according to the method, professional semantic accurate alignment is realized, the labor cost is greatly reduced, the corpus quality is reliable, the method can be expanded to similar fields, and efficient support is provided for field multi-mode large model fine adjustment and capability evaluation.

The application discloses a kind of for clinical diagnosis text automatic ICD-10 coding text matching model's construction method, device and application, comprising: first, by the coding history text is processed into priority matching set, with the coding result of coding history of previous years Analogy reasoning the coding result of unencoded clinical diagnosis text;Then, realize a kind of based on interactive text matching mode, in matching process, each character in two texts is interactively matched, uses the matching strategy of element-by-element subtraction and element-by-element multiplication through a neural network, retains more original information, obtains better matching effect, in aggregation process, using convolutional neural network, using the convolution kernel with same weight is merged feature to retain more original information, further improve the matching effect;Finally, data that cannot be matched with the coding history of previous years still can be secondarily matched with ICD-10 disease standard name, ensure the comprehensiveness of matching.

The invention provides a semantic segmentation-based paper evidence-based question and answer optimization method, which relates to the technical field of natural language processing, and comprises the following steps: performing rule segmentation on a paper text to obtain independent sentence units, and then performing recursive merging based on a dependency relationship between sentences and semantic relevancy to obtain a semantic segmentation result; forming a plurality of text blocks, and performing semantic vector representation and storage; a question input by a user is read and converted into a question semantic vector, and evidence fragments are extracted from the multiple text blocks based on semantic similarity and a text matching fusion strategy; and splicing the user input question and the evidence fragment into a unified input sequence, and inputting the unified input sequence into the generative model to generate an enhanced answer. Through the method, the technical problem that in the prior art, due to the fact that a traditional search engine is insufficient in semantic understanding and a mechanical fixed-length segmentation mode is adopted, evidence extraction is fragmented, and then the answer credibility is affected can be solved, and the accuracy and efficiency of paper evidence-based question answering are improved by fusing semantic segmentation.

This application discloses a medical question-answering method and related apparatus based on dynamic prompts and decoded knowledge editing, belonging to the field of artificial intelligence technology. The method first acquires a medical visual question-answering dataset containing medical images and question-answer pairs. Then, it pre-trains a medical visual language model using a contrastive learning framework, combining local contrast loss, global contrast loss, image-text matching loss, and masked language modeling loss to enhance the fine-grained and overall modal interaction between vision and text. Subsequently, dynamic prompts are added to the medical image and text features, eliminating the need for manually designed prompt words and adapting to multimodal feature differences. Next, the features with dynamic prompts are input into the pre-trained model to obtain initial predicted answers. Finally, through decoded knowledge editing, the output is corrected using a retrieval and verification mechanism from an external knowledge base, enhancing robustness in knowledge-intensive medical scenarios. Experimental results demonstrate that this method effectively improves the accuracy of medical visual question answering and performs excellently on various medical datasets.

The application provides a large model NL2SQL evaluation method and device based on database exploration, and relates to the technical field of data processing. The method constructs the semantic correctness evaluation process of NL2SQL as a multi-step decision process driven by a large model agent. The large model evaluation agent module generates a probing query by calling a read-only database execution tool module, iteratively updates the context according to the dynamically returned execution results, and outputs a semantic equivalence judgment. The present scheme changes the traditional "static text matching" evaluation paradigm to "dynamic execution verification". Through the closed-loop feedback mechanism of "proposing a hypothesis - executing verification - updating evidence", the evaluation bias caused by a single reference SQL is effectively solved, complex cases such as "accidental consistency of execution results but semantic error" or "inconsistency of execution results but semantic correctness" can be identified, and the execution accuracy of the evaluation and the robustness of the semantic discrimination are significantly improved.