43 results about "Character vector" patented technology

The invention relates to the technical field of natural language processing, in particular to a multi-granularity Chinese medical named entity recognition method and system for dynamic knowledge enhancement, and the method comprises the steps: constructing a medical term knowledge base containing a plurality of Chinese medical terms, and generating a term representation vector for each term; the method comprises the following steps: for an input Chinese medical text sequence, extracting a character initial embedding vector, fusing a word-level embedding vector of a word where the character initial embedding vector is located and an adjacent word-level embedding vector to obtain a multi-granularity feature fusion vector, and encoding to obtain a character vector containing context semantics; performing semantic alignment optimization based on the character vector and the term vector to obtain a term information enhanced character vector, and dynamically updating the medical term knowledge base according to the term information enhanced character vector; and taking the updated term vector as knowledge priori, performing multiple rounds of screening on the text candidate segments, and outputting a medical named entity recognition result. The method can effectively improve the recognition accuracy of nested entity, isomorphism and ambiguity terms, reduces boundary division errors and category confusion, and improves the reasoning efficiency.

The invention discloses an implementation method and system for AI and Agent agents to perform autonomous exploration and produce content, and relates to the technical field of AI and Agent agents, and the implementation method comprises the following steps: receiving an Agent identity file, an exploration target keyword, a narrative style parameter and an observation preference parameter configured by a user at a client, and generating Agent metadata containing character vectors and behavior tendencies; generating a virtual city model with a spatial topological structure based on GIS data, a POI map and historical and cultural corpora, and endowing buildings, roads and NPCs in a city with queriable semantic tags and interactive states to form a city-level semantic environment; by constructing a city-level semantic environment and a multi-agent collaborative exploration mechanism, autonomous perception, hierarchical decision and dynamic narrative generation of the AI Agent in a highly structured virtual city are realized, and by means of deep fusion of GIS data, POI atlas and historical and cultural corpora, a real geographic basis and cultural connotation are given to a virtual space, and immersion and logic coherence are improved.

The application provides an entity extraction method, device and equipment and a computer readable storage medium. The method comprises: obtaining at least one character vector and at least one extended word vector contained in a text to be extracted; the at least one extended word vector contains at least one preset entity vector; the at least one preset entity vector is the vector information of the entity corresponding to the text to be extracted in a preset entity dictionary; performing encoding and decoding transformation based on the at least one character vector and the at least one extended word vector to obtain at least one target entity corresponding to the text to be extracted; and the at least one target entity is used to realize natural language processing of the text to be extracted. Through the application, the efficiency of entity extraction can be improved on the basis of ensuring the accuracy of entity extraction.

The embodiment of the present disclosure discloses a text entity relationship analysis method, device, electronic equipment and readable storage medium. The text entity relationship analysis method comprises: a vector obtaining step, obtaining a vector combination comprising an entity vector corresponding to a character in a text, wherein the entity represented by the entity vector comprises at least one character; a vector splicing step, splicing the vectors in the vector combination for the character in the text to obtain a character splicing vector; a pooling step, performing pooling calculation on the character splicing vector to obtain an entity pooled vector; and an entity relationship determining step, splicing any two entity pooled vectors as an entity pair vector, and classifying the entity pair vector to determine the entity relationship.

The invention discloses a personalized emotion compensation method for lower limb walking intention perception of a stroke patient and a related device, and relates to the technical field of medical rehabilitation, the method comprises the following steps: synchronously collecting an electroencephalogram before a lower limb walking rehabilitation task of the stroke patient and a lower limb surface electromyogram signal in the task; preprocessing the electroencephalogram physiological signals and calculating dynamic brain function connection characteristics; preprocessing the electromyographic signals; inputting the brain function connection features into a TCN-LSTM model to obtain an initial emotion vector, and determining a comprehensive emotion vector in combination with an individual character vector converted by a psychological scale; time and amplitude scaling factors of the electromyographic signals are determined according to priori knowledge and a predefined optimization algorithm; inputting the preprocessed electromyographic signals into a convolutional network model to extract deep spatio-temporal features, performing scaling factor compensation to obtain compensated electromyographic features, inputting the compensated electromyographic features into a classifier to identify lower limb walking intention category output, and controlling rehabilitation equipment or feedback; the effect and efficiency of lower limb walking rehabilitation training of the stroke patient are effectively improved.

The present application relates to a kind of based on demonstrative word implementation financial field contract element extraction analysis method, including to document is divided into sentence, and text analysis is carried out;Initial vector representation of text character is obtained, aggregation demonstrative word is in the representation information in LSTM network layer, the similarity of demonstrative word and context other character vector is calculated, and weighted summation is carried out;Weighted information is converted into the probability distribution of each entity mark, and the optimal path is obtained by dynamic programming algorithm analysis, obtains the element entity information of text.The present application also relates to a kind of for realizing based on demonstrative word financial field contract element extraction analysis device, processor and its computer readable storage medium.The demonstrative word implementation financial field contract element extraction analysis method, device, processor and its computer readable storage medium based on the present application, compared with traditional method, guarantee element extraction accuracy, the whole time is far less than traditional method, and the efficiency of element extraction in contract can be greatly improved.

The invention provides a named entity recognition method for a high-speed railway technical improvement and overhaul project text, which comprises the following steps of: preprocessing unstructured data and semi-structured data in a picture and / or a text, and converting character-level positions of entity labels by adopting BIOES (Basic Input / Output Element Specification) labeling to obtain a technical improvement and overhaul data set; performing data enhancement on the technical improvement and overhaul data set by adopting a data enhancement strategy to obtain a data enhancement result; and inputting the text data subjected to data enhancement into an RBC fusion model, converting the text data into character vectors, carrying out a sequence labeling task to obtain a prediction label result of each character, and finally realizing an entity of a technical renovation and overhaul text through a large model prompt project method. Through preprocessing, data enhancement and an RBC fusion model, non-structured and semi-structured data in the high-speed rail technical improvement and overhaul project are efficiently identified, the problem that key information cannot be dynamically extracted through a traditional method is solved, and the method has the advantage that the information extraction efficiency and accuracy of the high-speed rail technical improvement and overhaul project are improved.

To extract metadata for highlights of a video stream from a card image embedded in the video stream.SOLUTION: Highlights may be segments of a video stream, such as a broadcast of a sporting event, that are of particular interest to one or more users. Card images embedded in video frames of the video stream are identified and processed to extract text. The text characters may be recognized by applying a machine-learned model trained with a set of characters extracted from card images embedded in sports television programming contents. A training set of character vectors may be pre-processed to maximize metric distance between the training set members. The text may be interpreted to be able to obtain the metadata. The metadata may be stored in association with a portion of the video stream. The metadata may provide information regarding the highlights, and may be presented concurrently with playback of the highlights.SELECTED DRAWING: Figure 4

The invention discloses a scene character recognition method based on single vector query decoding and related equipment. The method comprises the following steps: performing visual feature extraction on an image to obtain visual features; coding the character to be coded to obtain a character vector; performing global information fusion on the visual features and the character vectors to obtain context vectors; inputting the context vector and the visual feature into a single vector decoder for decoding to obtain a target decoded character; and if the characters in the image are not completely decoded, taking the target decoded character as a to-be-coded character, and returning to the step of coding the to-be-coded character by adopting the position sensing hash coding method to obtain the character vector until the characters in the image are completely decoded, thereby obtaining a scene character recognition result. According to the method, the recognition accuracy of the model on shielded and fuzzy samples can be improved, the problems of attention drift and calculation amount increase faced by a traditional autoregression decoding method can be relieved, and the method can be widely applied to the technical field of computer vision.

The application provides a term extraction method and device, electronic equipment and storage medium, the method comprises: performing word segmentation processing on input corpus of a to-be-extracted term to obtain a plurality of candidate words; for any candidate word, based on a word embedding language model, a word vector corresponding to the target candidate word is obtained; based on a character embedding language model, a character vector corresponding to each character in the target candidate word is obtained, and a sum vector of each character vector is further obtained; according to the similarity comparison result between the word vector and the sum vector, it is determined whether the target candidate word is extracted as a term; wherein the word embedding language model and the character embedding language model are both obtained by training a pre-trained language model based on sample corpus in the vertical industry to which the input corpus belongs. Thus, manual annotation of training data is not required, the term extraction cost is reduced, and accurate extraction of terms in vertical industry corpus can be realized.

The application provides a knowledge extraction method and system based on prompt learning, embedding obtained unstructured text data to obtain a cache vector value; embedding a homogeneous prompt string to obtain a homogeneous prompt symbol vector value, and embedding a heterogeneous prompt character string to obtain a heterogeneous prompt character vector value; splicing the cache vector value, the homogeneous prompt symbol vector value and the heterogeneous prompt character vector value to obtain a spliced vector, taking the spliced vector as a cache vector value of a pre-training language model; adopting a regular matching method to obtain structured data from text data generated by the pre-training language model; the application uses automatically encoded prompt characters, automatically learns a potential semantic representation of a label, and solves a knowledge extraction problem in a general way through a generative large-scale pre-training language model, thereby improving the precision and efficiency of knowledge extraction.

The disclosure generally relates to a computer-implemented method for training a machine learning model for text generation, the method comprising inputting text into the machine learning model; preprocessing the input text to obtain a plurality of character vector representations; encoding, using an encoder, each of the plurality of character vector representations to obtain a plurality of word vector representations; generating, using a backbone model, a plurality of predictive word vector representations based on the plurality of word vector representations; decoding, using a decoder, the plurality of predictive word vector representations to obtain a plurality of character-probabilities; and updating the machine learning model based the plurality of character-probabilities. The disclosure also relates to a computer implemented method for generating text, a corresponding device, system and computer program.

The application provides a network request risk detection method, system and server, and relates to the field of network request risk detection. The method performs feature fusion on an injection keyword matching result, a network request parameter statistical result and a network request character vectorization result corresponding to to-be-detected data, dynamically optimizes continuous features in the fused feature data, and realizes accurate detection of SQL injection risks in a network request process through a lightweight classifier.

This invention discloses a work order classification method, system, device, and medium based on a multi-classification model, relating to the field of work order classification technology. The method includes: acquiring original work order text and forming a labeled sample set through expert labeling; statistically analyzing the label distribution and merging or data augmenting categories below a threshold to obtain a balanced labeled dataset; cleaning to remove non-semantic information to obtain a clean text dataset; training a static character vector model and a pre-trained dynamic character vector model to obtain static and dynamic vector weights; extracting and concatenating static and dynamic features to generate a fused feature matrix; hierarchically sampling the fused feature matrix into training and testing sets, inputting it into a Bi-LSTM-Attention-CNN multi-classification network for training to obtain a work order classification model for classifying new work orders. This invention can replace manual work order automatic labeling for tens of thousands of work orders, saving 2000 hours annually and directly providing data support for product iteration.

A method for extracting Chinese financial evaluation units based on inter-character relationships includes: encoding Chinese financial text sentences at the encoding layer to obtain character vectors; feeding the character vectors and the inter-character relationship matrix into a graph convolutional neural network model, enhancing the Chinese character relationships between evaluation elements on the character vectors, and encoding the dependency relationships; constructing a grid of character pair relationships in the evaluation text, and differentiating the words on the rows and columns of the grid using two unit convolution kernels; performing convolution operations using image convolution kernels of different sizes in the row and column directions; for elements in the grid, fusing their corresponding word codes on the rows and columns and the dependency relationship codes between the two words, determining the element's label by decoding, and completing the extraction of the evaluation unit. The present invention can solve the problems of insufficient utilization of Chinese character relationships between evaluation elements, failure to distinguish different evaluation element types when representing grid elements, and insufficient utilization of Chinese character relationships within evaluation elements.

The application provides a semantic recognition method and device, the semantic recognition method comprises: obtaining N first texts and a second text; each first text comprises a first keyword in a preset word table, and N is a positive integer greater than 1; performing first text processing on each first text through a preset text processing model to obtain M character vectors corresponding to the first keyword in each first text, and M is a positive integer greater than 1; determining a first word vector corresponding to each first text according to the M character vectors; and performing semantic recognition on the second text according to the first word vector corresponding to each first text.

The application belongs to the technical field of computer application, and particularly relates to a college major attribute entity extraction method based on a difference attention mechanism. The method takes a character vector, an external vocabulary vector and an entity text vector as keys, values and queries of the difference attention mechanism, effectively integrates the difference of external knowledge into the original text, improves the personalized representation of the character vector in different contexts, realizes better fusion, simultaneously, utilizes a CRF model to perform structure output, and further realizes the college major attribute entity extraction, attributes are assigned to the college majors, so that the candidates can conveniently select.

The invention discloses an ancient book entity recognition method and system, a medium and equipment, and relates to the technical field of ancient book named entity recognition, and the method comprises the steps: obtaining a to-be-recognized ancient book text statement; the method comprises the steps of converting an input ancient book text statement into an initial word vector, determining corresponding character features and syntactic features according to the initial word vector, and obtaining a text with context information according to the character features and the syntactic features; wherein the initial word vector is the superposition of a character vector, a fragment vector and a position vector; feature extraction and splicing are carried out on the text from the forward direction and the reverse direction, splicing features are determined, the entities and punctuations are recognized according to the splicing features, and corresponding ancient custom entity recognition results are obtained; wherein the identification result comprises an identification tag sequence and a punctuation tag sequence.

This disclosure presents embodiments of an information extraction method, apparatus, device, and computer-readable medium. One specific implementation of the method includes: acquiring target text; encoding each character in the target text to generate character vectors, obtaining a character vector sequence; determining a target probability value group corresponding to each character vector in the character vector sequence, obtaining a target probability value group sequence; generating an object vector sequence set based on the target probability value group sequence and a tag set; generating a tag sequence set based on the object vector sequence set and an object transition matrix set; and extracting object information from the target text corresponding to each tag sequence in the tag sequence set, obtaining an object information set. This implementation enables information extraction from text with overlapping information, providing convenience for applications such as text analysis.

Methods, apparatuses and computer programs for training a machine learning model to generate text and for generating text using the trained machine learning model are disclosed. The present disclosure generally relates to a computer-implemented method for training a machine learning model for text generation, the method comprising: inputting text into the machine learning model; preprocessing the input text to obtain a plurality of character vector representations; encoding each of the plurality of character vector representations using an encoder to obtain a plurality of word vector representations; generating a plurality of predicted word vector representations based on the plurality of word vector representations using a trunk model; decoding the plurality of predicted word vector representations using a decoder to obtain a plurality of character probabilities; and updating the machine learning model based on the plurality of character probabilities. The disclosure also relates to a computer-implemented method for generating text, a corresponding device, a system and a computer program.

This application discloses an image processing method, apparatus, device, and computer-readable storage medium, belonging to the field of computer technology. The method includes: obtaining character vectors corresponding to the characters in the text content to be processed and key information vectors corresponding to the key information in the text content; determining initial image features corresponding to the text content; adjusting the initial image features based on the character vectors and key information vectors to obtain target image features corresponding to the text content; and obtaining the target image corresponding to the text content based on the target image features. This method considers relatively comprehensive information, and the obtained target image can fully cover the text in the text content, resulting in a high degree of matching between the target image and the text content.

An information processing method, device, equipment, storage medium and program product are disclosed, the method comprising: obtaining a to-be-detected target sentence of a user, the to-be-detected target sentence being composed of multiple characters; determining corrected characters according to the multiple characters through an error correction model; the error correction model being obtained by training a pre-trained language model, a semantic aggregation layer network and a classifier, the pre-trained language model being used to generate character vectors, the semantic aggregation layer network being used to generate error correction vectors according to the character vectors, and the classifier being used to generate the corrected characters according to the error correction vectors; wherein the corrected characters compose a corrected sentence, which is used to support voice interaction with the user. The present application can solve the problem of complex processing and low accuracy of error detection in the prior art voice error correction method.

The present disclosure provides a traditional Chinese medical record named entity recognition method based on a multi-head attention mechanism, comprising the following steps: obtaining text data of a traditional Chinese medical record; combining character vectors and word vectors in the obtained text data and sending them into a Bi-GRU neural network for feature extraction to obtain global features; using a multi-head attention mechanism to provide potential semantic information for the character vectors and extract local features; inputting the global features and the local features into a conditional random field layer to obtain a named entity sequence labeling result of the text data; the present disclosure does not require a word segmentation operation, combines the features of characters and words to form a joint feature, controls the weights of the characters and the words using a hyperparameter, inputs the joint feature into an embedding layer, and adds spatial attention in the Bi-GRU layer to make up for the deficiency in extracting effective features, greatly improving the accuracy of entity recognition.

The invention discloses a traffic accident address analysis method and device, equipment and a medium, and relates to the technical field of artificial intelligence. An address text is analyzed into a plurality of address elements, and an address element space relation directed graph is constructed based on the address elements and corresponding space semantic confidence degrees; the spatial semantic relation in the address elements is quantized, so that the road boundary rule is captured; the character vectors and the corresponding spatial semantic feature vectors are spliced according to the distributed weights to obtain a plurality of preliminary feature vectors so as to achieve conversion and unification of boundary rules and vector spaces, then the characters and the spatial vocabularies are matched based on the correlation degree, the converted and unified characters and the spatial vocabularies are organically matched, and therefore the matching efficiency of the characters and the spatial vocabularies is improved. Therefore, organic unification of traditional discrete rule boundaries and continuous address space features is realized, the analysis efficiency of Chinese addresses in traffic accidents is improved, and a basis is provided for traffic accident processing.

The application discloses a character image infringement early warning method and device, equipment and medium, and relates to the technical field of computer vision, which comprises the following steps: frame extraction is performed on a to-be-detected video, character detection is performed on the extracted target frame, and detection information is generated; a to-be-detected vector and a to-be-detected character visual constituent element are generated based on the detection information, the to-be-detected vector is matched with a registered character, and a candidate registered character is determined; a target score is determined based on the registered character vector and the to-be-detected vector, and an element combination constraint score of the to-be-detected character image and the candidate registered character is determined; a comprehensive credibility score is determined based on the target score and the element combination constraint score, the comprehensive credibility score is aggregated, and an aggregated score is obtained; a candidate evidence package is constructed according to the aggregated score, the target score, the element combination constraint score and a registered reference graph; the candidate evidence package is reviewed, and an infringement early warning result is determined according to the aggregated score and a structured review result. The character image after modification can be accurately identified.

The embodiment of the application discloses an information extraction model training method, an information extraction method, an information extraction device, an information extraction equipment and a medium. The information extraction model training method comprises the following steps: obtaining information extraction training data; inputting the information extraction training data into a label vector output submodel of an information extraction model for label identification to obtain a label representation vector of the information extraction training data; inputting the information extraction training data into a character vector output submodel of the information extraction model for character identification to obtain a character representation vector of the information extraction training data; calculating similarity data of the label representation vector and the character representation vector; and inputting the similarity data into a classifier model of the information extraction model for classification identification to obtain an information classification result of the information extraction training data. The technical scheme of the embodiment of the application can improve the training efficiency and model precision of the information extraction model, and further improve the efficiency and accuracy of information extraction.

The application provides a Chinese event detection method based on a part-of-speech attention mechanism, which is based on a public data set. First, the sentence is divided into words using a word segmentation tool on the data set, and then the part-of-speech sequence of the sentence is obtained using a part-of-speech tagging tool. The part-of-speech sequence of the sentence is input into a CBOW model to obtain a pre-trained part-of-speech vector, so as to learn the fixed collocation information between words, such as "suffered an injury" as a "verb + adverb + noun" structure. Then, the part-of-speech vector, the word vector and the character vector are used to extract the word-level information and the character-level information of the sentence, respectively. When extracting features, a convolutional neural network is used to extract features on the word matrix, the character matrix and the part-of-speech matrix of the sentence. Then, the part-of-speech features are used to calculate the attention score, which is used to assist the model to focus on the verb when calculating; the part-of-speech attention is provided, and after the part-of-speech features are added, the accuracy and efficiency of the model in the trigger word extraction and event type classification tasks are higher.

The invention provides a network request risk detection method and system and a server, and relates to the field of network request risk detection.The method comprises the steps that feature fusion is conducted on an injection keyword matching result, a network request parameter statistical result and a network request character vectorization result corresponding to data to be detected; and after continuous features in the fused feature data are dynamically optimized, accurate detection on the SQL injection risk in the network request process is realized through a lightweight classifier.

This information retrieval system provides an answer corresponding to a question using a large language model. A context information retrieving unit retrieves a document database with a characteristic vector of the question and thereby acquires as context information a text group that a similarity level between the characteristic vector and a characteristic vector of the text group satisfies a predetermined condition. In the document database, character vectors and page numbers of text groups obtained by dividing a document are registered. A prompt generating unit generates a prompt that includes the question and the context information. An answer acquiring unit acquires an answer corresponding to the prompt using a large language model. An answer outputting unit outputs as an answer corresponding to the question the context information and a page number associated with the context information together with the answer corresponding to the prompt.

The application discloses a Chinese speech offline recognition method and system based on a Transformer, and the Chinese speech offline recognition method comprises the following steps: inputting speech, extracting Mel frequency cepstrum coefficients in the speech and converting the Mel frequency cepstrum coefficients into an input sequence, and inputting a neural network model; an encoder of the neural network model encodes the input sequence to obtain a character vector; a decoder of the neural network model decodes the character vector to obtain output features; the output features are fused to obtain integrated features; the integrated features are subjected to softmax processing to obtain a predicted character set; and the predicted character set is subjected to weight calculation based on a word group dictionary to obtain accurate characters. In the application, features in the decoder of the Transformer are fused to the top layer, the generalization features of an audio signal are better extracted, and weight calculation is performed through the word group dictionary, so that the problem of homophonic words can be effectively solved.