22 results about "Word identification" patented technology

The invention discloses a training method for a high-performance wrongly written character recognition model, and relates to the field of model training, and the method comprises the steps: obtaining character and word associated information, and constructing a character and word knowledge graph; extracting word multi-dimensional feature vectors, and performing sample expansion on the primary sample set to obtain a word and sentence sample set; constructing a wrongly written character recognition model, wherein the wrongly written character recognition model comprises a primary recognition main model and a primary recognition sub-model; completing iterative training of the primary identification main model to obtain a wrongly written character identification main model; outputting a wrongly written character recognition result set through the wrongly written character recognition main model; and training the primary recognition sub-model by using the wrongly written character recognition result set and the word and sentence sample set to obtain a wrongly written character recognition sub-model. The wrongly written character recognition precision and recognition efficiency of the wrongly written character recognition model can be effectively improved.

The invention discloses a service-adaptive satellite-borne non-uniform multicarrier demodulation device, and belongs to the field of digital signal processing. The method comprises the following steps of: firstly, performing orthogonal down-conversion and resampling filtering processing on a digital intermediate frequency signal sampled by an ADC (Analog to Digital Converter), processing the digital intermediate frequency signal by a non-uniform channelization splitter, and then serially outputting the digital intermediate frequency signal to a spread spectrum ratio self-adaptive variable burst capturer according to fixed quadruple sampling to capture a leading head and identify the spread spectrum ratio; the method comprises the following steps: completing symbol synchronization and de-spreading processing on an input signal according to a spread spectrum ratio, then completing carrier coarse synchronization and pattern word identification, completing modulation mode identification and carrier synchronization according to a pattern word, performing channel decoding through a multi-mode decoder to recover to an information bit stream, and outputting the information bit stream. And the information bit stream is sent to a service frame recombination processor for search identification and recombination splicing, and output is carried out after whole-frame CRC check is carried out. The method has the advantages of flexibility in use, low resource occupation and capability of realizing adaptive demodulation of variable-rate, multi-modulation-mode and multi-coding-mode service transmission signals without parameter configuration.

The application designs a sensitive data prediction method based on user editing track in a collaborative editing scene, relates to the field of data security, and comprises the following steps: obtaining user historical text data, screening target text containing sensitive data, performing theme division, word segmentation, extracting keywords and position information to establish a text editing track; for each keyword of the text keyword set, a context window mechanism is adopted to select a sensitive word set and divide the sensitive word security level; according to the keyword editing clue, the theme label and the sensitive word set with the level division, a graph attention network is used to construct a user track model based on keywords; inputting the editing process text, identifying the keywords and labels, combining the received person role identity level, and inputting the track graph model to obtain a predicted sensitive word set. The application creates a model based on the user editing track, can accelerate the identification efficiency of sensitive data, reduce the risk of leakage, and improve the sensitive word identification accuracy under different text themes.

The current invention relates to an imagined, performed, mouthed or attempted speech fragment identification and semantic reconstruction. The method comprising the steps of: receiving, via a brain data capturing element, a brain data fragment containing speech; and searching for a command comprising at least one word, contained in said brain data fragment, said search being carried out based on a database comprising data related to at least one predetermined command, after the step of receiving a brain data fragment containing speech, an identification window having a starting-point, a duration and an end-point is selected in the brain data fragment, and within which identification window a word-specific word identification model is used in order to identify at least one word contained in at least one of the predetermined commands. In a second aspect, the invention relates to a system for controlling an electromechanical device.

The invention discloses a three-dimensional-based component label automatic generation method and device and a medium, belongs to the technical field of three-dimensional views, and aims to solve the technical problems that in an existing heating and ventilation three-dimensional view, the manual label efficiency is low, labels are prone to overlapping and confusion, and the readability and construction accuracy of drawings are affected. The method comprises the following steps: performing view frame selection processing on a three-dimensional building component to obtain a locked three-dimensional component view; carrying out keyword identification and character labeling processing of related component family categories on the three-dimensional component view to obtain initial component labeling information; performing labeling area limitation on a single three-dimensional building component in the three-dimensional component view to obtain a to-be-labeled view area; performing local optimal scoring processing on the labeling position in the target labeling division area to obtain a local optimal labeling position; and performing related global annotation optimal position control on each local optimal annotation position in the three-dimensional building component to obtain final annotation content position information.

The invention discloses wearable interaction hardware equipment, and relates to the technical field of man-machine interaction. S1, equipment wearing and setting; s2, reporting contact information; s3, automatically replying the voice message; s4, performing voice reply by the user; and S5, traffic light pause state notification. According to the wearable interaction hardware equipment, a user does not need to hold a mobile phone to check or edit messages, distracted driving is avoided through voice interaction of a Bluetooth headset, the traffic accident risk is reduced, meanwhile, road congestion caused by distraction of a driver is reduced, and timely transmission of emergency messages and delayed processing of non-emergency messages are ensured through emergency feature word recognition; not only are important cooperation or help information not missed, but also normal driving is not interfered, the non-emergency message is automatically replied and the optimal contact time is informed, the communication cost of a user is reduced, and the non-emergency message is broadcasted and replied on the premise of ensuring sufficient time in combination with a traffic light pause state, so that the message processing efficiency is improved, and message accumulation is avoided.

The invention discloses a sensitive word evolution method, a sensitive word recognition method, equipment, a medium and a product. The method comprises the steps of obtaining a sensitive word seed set; initializing an evolution diagram according to the sensitive word seed set, and initializing an evolvable node pool according to the evolution diagram; determining an evolution factor corresponding to each evolvable node in the evolvable node pool, determining an evolution priority of the evolvable nodes according to the evolution factors, screening out nodes to be evolved according to the evolution priority of each evolvable node, and determining a target evolution operator of each node to be evolved; evolving the node to be evolved according to the target evolution operator, and updating the evolution diagram and the evolvable node pool according to a variant obtained through evolution; judging whether an evolution stop condition is met or not, and if yes, generating a sensitive word variant set according to variants obtained through evolution; and if not, returning to execute the step of determining the evolution factor, so that the problem that the sensitive words cannot be systematically generated is solved, a large number of sensitive words are automatically generated, and the generation efficiency of the sensitive words is improved.

The application discloses a document specification review method and system, comprising: obtaining a to-be-reviewed document and preprocessing, calling a pre-constructed professional vocabulary error correction model of a corresponding field to perform error word identification, and obtaining an error word identification result; judging whether the pre-processed to-be-reviewed document contains a reference standard, if the reference standard is not identified, generating a reference standard missing prompt; if the reference standard is identified, extracting information of the reference standard, comparing the identified reference standard information with a pre-constructed standard specification database, and judging whether the reference standard in the to-be-reviewed document is a valid version; if the reference standard in the to-be-reviewed document is a valid version, performing content consistency comparison between the to-be-reviewed document and the pre-constructed standard specification knowledge base, and judging whether there is a description error. Through the method, error words and standard specifications can be identified, and the identification accuracy and efficiency are improved.

Language models require resource-intensive training. As a result, retraining language models happens very infrequently even though new or custom words are needed at a much faster rate. By deploying a custom set of words, speech recognition may be performed using a previously trained language model and augmented with entries in a custom word list. A probability map is created for each token position predicted by the model. Next, potential positions for custom words are identified by calculating the probability ratio between the token selected by the model and the custom word token. The probability ratios are summed for the first, second, and last tokens of the custom words, and if the sum falls below a certain threshold, the position is recorded. Next, the word or words starting at the recorded position are identified and, using string comparison metrics, are determined the most likely candidates for replacement.

The invention discloses a data processing method and device, a storage medium and electronic equipment. The method relates to the technical field of artificial intelligence, and comprises the following steps: carrying out sensitive word identification on data in a target data set to obtain a data label of each data in the target data set; first data in the target data set is identified through the target model, an identification result is obtained, the first data is composed of data with a second label, and the identification result is used for representing whether the first data contains sensitive data or not; and desensitizing second data in the target data set to obtain a processed target data set, the second data being sensitive data. Through the method and the device, the technical problem of relatively low data security caused by low sensitive data recognition accuracy in related technologies is solved.

The invention discloses a news propaganda communication management method and system and a program product. The method comprises the following steps: extracting a subject term set of news text data; secondly, generating a word feature vector and a named entity feature vector of the subject term set, and performing clustering processing on the subject term set based on the word feature vector and the named entity feature vector, so that subject terms with similar word features and named entities are clustered into a class cluster; then, the hot words in the news topic class clusters are recognized by calculating the word popularity of each topic word in each news topic class cluster, so that the hot words in each class cluster are obtained, and the hot word recognition mode based on the hot words and the word popularity can ensure that the pushed hot words have higher attention and news value; therefore, the hot words are generated through a natural language processing technology, and compared with traditional manual topic selection, the time cost is reduced, the efficiency is improved, the subjective experience influence is reduced, and the hit rate of the hot words is guaranteed.

The invention relates to a transfer learning-based special field Chinese word segmentation method and system, and aims to solve the problems that culture load words in international education field texts are difficult to recognize, the adaptability of a general model field is poor, and semantic consistency optimization is insufficient. Enhancing cultural load word representation in combination with an attention mechanism; eliminating domain differences by using adversarial training and feature distribution alignment technologies, and generating domain invariant features; fusing sequence labeling and boundary detection tasks to optimize word segmentation boundaries; and finally, dynamically adjusting the segmentation path based on the semantic similarity score. According to the method, the word segmentation accuracy of the composite terms in the education field is remarkably improved, the problems of wrong segmentation of culture load words, reduction of field migration performance and lack of semantic coherence are solved, and high-robustness word segmentation support is provided for an international education text processing system.

Embodiments of the present application relate to a kind of interference word identification method, device, electronic equipment and storage medium, the method includes: obtaining to be identified text, the to be identified text is carried out word segmentation processing, obtain the word segmentation set of the to be identified text;For each word segmentation in the word segmentation set, obtain the target text after the to be identified text deletes the word segmentation;Determine the semantic difference degree between the to be identified text and the target text and user behavior difference degree;According to the user behavior difference degree and the semantic difference degree, identify whether the word segmentation is interference word.Thereby, interference word identification can be carried out from two aspects of semantic difference degree and user behavior difference degree, improve the accuracy of interference word identification of to be identified text.

The invention relates to the technical field of electric energy meter production detection, and provides an electric energy meter detection data full-process tracking method based on status word identification, and the method comprises the steps: setting a unique identity identification for an electric energy meter, and setting a status data block which can be encrypted and edited in the electric energy meter; inputting the unique identity label of the electric energy meter into an internet platform, and synchronously uploading the state of the current link to the internet platform for backup when the state of the electric energy meter is written; inputting the market order parameters into an internet platform, and matching the market order parameters with the unique identities of the electric energy meters and the state data blocks of the electric energy meters through the internet platform when the electric energy meters are delivered out of a warehouse; and when the electric energy meter enters a detection link, the detection equipment automatically reads the asset number of the electric energy meter, generates a corresponding status word according to the current link and sends the status word to the detection equipment, and after the detection equipment completes detection, the detection data and the status word are packaged and uploaded to the database, so that real-time binding of the detection data and the status word is realized.

The embodiment of the invention discloses a ship controller off-line voice instruction control method and system. The method comprises the following steps: acquiring a ship environment voice signal, and converting the ship environment voice signal into a first digital audio stream; performing audio front-end enhancement processing on the first digital audio stream; performing wake-up word local offline detection on the first audio feature based on a local lightweight neural network model; acquiring a second audio feature, and performing local offline identification on the second audio feature through a local command word identification model; analyzing the natural language instruction into a structured control instruction based on a predefined ship control instruction mapping table; and based on a system integration mode of multi-mode instruction forwarding, sending the structured control instruction to a ship controller through a wired serial port communication or Bluetooth low-power-consumption wireless communication mode. According to the embodiment of the invention, ship background noise and echo interference can be effectively suppressed, and high recognition accuracy can be kept.

The application discloses a chapter-level event extraction method and device, equipment and a storage medium. For a target chapter carrying title information, event parameter extraction is directly performed without trigger word identification. An event extraction result is composed of parameter values of various types of event parameters, event names and event types, and is more in line with chapter characteristics. Since trigger word identification is omitted, the entire processing procedure is simpler. In the event parameter extraction process, the target chapter is divided into sentences, and parameter values of various types of event parameters in each sentence are extracted in sequence according to an event parameter type template. Since event parameter extraction is performed at the sentence level, the marked event parameters can be prevented from being chaotic. For each type of event parameter, the parameter values of the type of event parameter extracted in each sentence are integrated to obtain integrated parameter values. The integrity of the parameter values of various types of event parameters extracted at the chapter level is ensured, and the problem of chaotic marked event parameters is avoided.

The application discloses a marketing mobile terminal dynamic behavior compliance monitoring method and system, belongs to the technical field of information transmission monitoring, and aims to solve the problems of insufficient non-text information processing, sensitive word determination redundancy, incomplete coverage of illegal expression variants and low data reuse rate of existing monitoring. The method comprises the following steps: capturing dynamic behavior information, constructing a composite retrieval code retrieval history library, and judging whether sensitive word identification has been performed; if not, a standardized word segmentation sequence is generated; word segmentation comparison is performed to determine matched word segmentation and unmatched word segmentation, the matched word segmentation history synonym set is called, the unmatched word segmentation core semantic screening avoidance expression is positioned, the synonym set thereof is generated, and the synonym data set is constructed; the matched word segmentation history judgment result is called, the synonym sequence algorithm probability is established, the sensitive word is judged in combination with the significant proportion, the processing flow is executed, and the history library and the word segmentation library are updated. The application realizes efficient and accurate compliance monitoring, improves the data reuse rate, and guarantees the effectiveness of the monitoring closed loop.

The application belongs to the technical field of communication signal processing, and provides a frame synchronization word identification method, device, equipment, medium and product, which are used for improving the accuracy and adaptability of frame synchronization word identification. The method inputs at least one data signal block obtained through a sliding time window into a first network model for extracting local dependence features of time sequence information, obtains at least one local time sequence dependence feature, inputs the at least one local time sequence dependence feature into a second network model, and obtains a recognition result output by the second network model for indicating a signal data frame synchronization word, so as to improve the accuracy and adaptability of frame synchronization word identification.

The invention discloses an ACG-oriented daily source word special item recognition and corpus construction method and system, and the method comprises the steps: collecting ACG full-scene texts, carrying out the processing of denoising, word segmentation and the like, carrying out the hierarchical storage according to time, building a high-quality traceable data basis, and providing a support for the subsequent analysis. Through N-gram statistics and solidification degree and context entropy calculation, high-frequency and compact combinations with free boundaries are screened, potential daily source words are purified, and the verification workload is reduced. Daily source attributes are confirmed through stepped verification, classified storage is carried out according to language irreplaceability after multi-dimensional labeling, and a structured corpus is constructed. Propagation rules and semantic features are mined through time sequence and network analysis, and corpora and models are dynamically updated. All the steps are progressive layer by layer, the front step is the rear step for input, and a closed loop of data preparation, screening, verification and analysis optimization is formed. Technical support is provided for ACG daily source word recognition, corpus construction and rule mining.

The application discloses a kind of natural language question in intelligent question answering system-oriented focus word identification method, it is related to natural language question answering technical field.The application proposes the method for identifying focus word in natural language question, so that question answering system can more accurately understand the user's focus point;And propose a prefix tree structure dominated by decision item, based on the prefix tree, further introduce an algorithm for mining strong focus association rules to identify focus word, the algorithm is more efficient than the classical association rule mining algorithm Apriori;And propose a kind of inverted index for strong focus association rules, based on the inverted index, further introduce an algorithm for identifying focus word, the algorithm is more efficient than sequential search;And define focus item set, frequent focus item set, focus association rule and strong focus association rule, to better express information related to focus word.

The utility model discloses an audio frequency detects early warning module, many people review monitoring system and device, wherein the module includes voice collection subassembly, warning subassembly, closing subassembly and processing assembly, processing assembly includes speech recognition unit, and voice collection subassembly and warning subassembly all are connected with speech recognition unit, and closing subassembly is connected with warning subassembly, voice collection subassembly is used for collecting sound data and transmission to speech recognition unit, and speech recognition unit is used for the sensitive word identification in sound data and exports warning signal to warning subassembly and starts warning, and closing subassembly is used for controlling warning subassembly and stopping warning. The utility model realizes the risk reminding and warning in time in the review process, and it is favorable for reducing the occurrence of illegal operation.

The application relates to the technical field of electric energy meter production and detection, and provides an electric energy meter detection data whole-process tracking method based on state word identification, which comprises the following steps: setting a unique identity for an electric energy meter, and setting an encryptable and editable state data block in the electric energy meter; entering the unique identity of the electric energy meter into an internet platform; when the state of the electric energy meter is written, synchronously uploading the state of the current link to the internet platform for backup; entering market order parameters into the internet platform; when the electric energy meter is delivered, matching the market order parameters with the unique identity of the electric energy meter and the state data block of the electric energy meter through the internet platform; when the electric energy meter enters a detection link, the detection equipment automatically reads the asset number of the electric energy meter, generates a corresponding state word according to the current link and sends the state word to the detection equipment, and after the detection equipment completes detection, the detection data and the state word are packaged and uploaded to a database, so that real-time binding of the detection data and the state word is realized.