62 results about "Language semantics" patented technology

The invention discloses a translation model establishing method and system. The translation model establishing method comprises the following steps: respectively generating a regular alignment table, a word semantic vector table and a phrase table according to alignment information of a double-language parallel corpus, subsequently generating a source language phrase semantic vector table of a source language semantic space and a target language phrase semantic vector table of a target language semantic space by using the word semantic vector table and the phrase table, and finally training by using phrase semantic vector tables of different semantic spaces, thereby generating a translation model integrated with semantic information. The result shows that phrase semantic information can be integrated in statistic machine translation, the research shows that the relevance of words or phrases to context words or phrases can be reflected in the semantic information, and compared with a conventional translation method based on words or phrases, the translation model is relatively high in translation quality after the phrase semantic information is integrated, so that the translation property of the statistic machine translation is further improved as compared with that of the prior art.

A concept-based back translation system includes a target language semantic parser module, a source language semantic parser module, a bi-directional machine translation module, a relevancy judging module, and a back translation display module.

The present invention discloses a semantics-based cross-language patent novelty predetermining analysis method. The method comprises the steps of reading classified information of Wikipedia, extracting keywords, synonyms and related words of an article title, a disambiguation page and a redirection page, and building a cross-language semantic network structure; extracting Chinese keywords from a to-be-examined Chinese patent, reading related words of the Chinese patent from a cross-language knowledge bank, combining and expanding the multiple Chinese keywords and the related words, so as to form multiple groups of Chinese inquire words, translating the Chinese inquire words into corresponding English inquire words for retrieving, so as to obtain multiple similar English files; and extracting English keywords, performing similarity calculation, predetermining the novelty, building a tree map through hierarchical clustering, and outputting the most similar English files. Through adoption of the method, a cross-language comparison policy is made, patent writing and examination efficiency can be improved greatly, patent declaration and approval time is shortened, patent declaration and examination efficiency is improved, and the method has high promotion value.

The invention allows a functional object in a dynamic language to be efficiently used as both a directly-invoked function and as a method in a generic function. This allows a language to combine the two concepts, simplifying the language semantics. As a preferred embodiment, a method object when created comprises as contiguous fields a header, a method-information section, a function prolog, and the method body. When called as a component of a generic function, one entry point is at the method body or at the method-information field. When called directly, another entry point exists at the function prolog field.

The invention provides a low-resource language entity extraction method based on bilingual word vectors. Semantic features of languages are considered in a low-resource language entity extraction task, and the problem of unsupervised learning of low-resource language entity extraction is solved. The method comprises the following three stages: step (1), constructing bilingual word vectors based ona comparable corpus; (2) constructing a source language entity extraction model; and (3) constructing a target language entity extraction model. Compared with the prior art, the reinforcement learning and the bilingual word vector are introduced into the low-resource language entity extraction task for the first time, and the problem that the low-resource language lacks an entity extraction annotation corpus is solved. Through the bilingual word vectors, word meaning characteristics of the cross-language text are effectively represented, and the problems that low-resource language semantic information is insufficient and semantic information cannot be directly transferred among languages in the model transferring process are solved. Meanwhile, the reinforcement learning thought is adopted, and unsupervised learning of the low-resource language entity extraction task is achieved.

The invention discloses a natural language semantic characterization method based on an attention mechanism, which comprises the following steps of: introducing a multi-layer attention mechanism and aposition relation matrix into semantic relation feature extraction so as to obtain more semantic information in language semantic characterization of a natural statement; In this way, the deep learning technology and the attention mechanism are fully combined to share the multi-layer semantic features and the semantic position information features representing the natural language, the semantic features of the language are fused, the algorithm performance is improved, meanwhile, the semantic universality is improved, and the method can be used for various natural language processing tasks.

The invention discloses a management method and a management system for resource information. The method comprises the following steps of: creating or utilizing attribute information in an attribute information library, wherein the attribute information comprises the semantic description information of attributes and semantic logic relationship information used for classification and retrieval processes; and creating a resource library and a directory library according to the attribute information in the attribute information library, wherein the resource library is used for storing resource identification information corresponding to information resources to be managed and the semantic logic relationship information of reference corresponding to the information resources in the attribute information library, and the directory library is used for storing directory identification information corresponding to directories created by users and the semantic logic relationship information of reference corresponding to the directories in the attribute information library. Private information is protected by separating semantic information and the semantic logic relationship information, a semantic searching algorithm is optimized by combining the structural processing of the semantic logic relationship information, and retrieval efficiency is improved; and in addition, the comprehensive processing of common attributes is combined to realize cross-language semantic retrieval.

The invention provides a semantic processing method for a semantic comprehension model. The semantic processing method comprises the steps of obtaining voice instruction information and noise information of a vehicle-mounted environment; in response to the voice instruction information, converting the voice instruction into corresponding recognizable text information according to the noise information of the vehicle-mounted environment; and triggering a corresponding business process according to an object matched with the hidden variable of the word level and a task field corresponding to thehidden variable of the word level so as to complete a task corresponding to the voice instruction information. The invention further provides a processing device for the language semantic comprehension model and a storage medium. According to the method, the training precision and the training speed of the semantic comprehension model can be improved, so that the semantic comprehension model canadapt to a vehicle-mounted full duplex use scene, the influence of environmental noise on the semantic comprehension model is avoided, and the accuracy of semantic comprehension in a vehicle-mounted full duplex environment is improved.

The invention discloses an AR remote control device, an intelligent home remote control system and a method, which relate to the AR field. The AR remote control device comprises an AR display device for capturing images of the intelligent home; an instruction control device receives a control instruction and sends a control instruction encoding signal to the smart home according to the control instruction. The remote controller based on AR technology of the invention makes any user of the remote controller feel concise and convenient, especially for the patient who loses the language ability,which is a gospel artifact. With the help of language semantic recognition technology, the elderly and children users feel warmer.

The invention provides a dynamic sign language semantic recognition system and method based on a depth image. According to the system and the method, depth image video information of an operator is acquired; the video information is processed to obtain hand joint information, sign language words are analyzed through the joint information; inputting each word into a semantic analysis model; and judging whether the semantic expression is complete or not, and directly outputting or converting control commands to other control units when the intention expression is complete, so that the sign language actions are translated into characters, a machine or an operating system is controlled, and hearing impairment people can be better integrated into social life.

Aiming at shortcomings of scene construction methods and systems of prior social tools, a scene robot is used to carry out natural language understanding (NLU) according to conversation contents between conversation subjects to obtain conversation language-environment information, then a database is associated for acquiring data, medias\or a virtual reality technology\or an augmented reality technology is used to generate a scene, and transition from a language environment to the scene is realized; and the generated scene is fused with a conversation\or exists in time and space in an associated and dependent manner, so that the conversation scene of a social tool is constructed, understanding of people on social-language semantics is enhanced, and the expressive force and the sentiment of the conversation contents are improved.

The invention provides a power grid dispatching professional language semantic relation extraction method and device and electronic equipment, and the method comprises the steps: collecting a dispatching history corpus generated in the operation process of a power grid; extracting vocabularies from each corpus text in the scheduling history corpus to obtain a plurality of vocabularies contained inthe corpus text, and constructing a vocabulary vector corresponding to each vocabulary according to the arrangement sequence of the plurality of vocabularies in the corpus text; Constructing a semantic vector corresponding to each vocabulary based on the vocabulary vector corresponding to each vocabulary and a preset neural network model; and calculating semantic similarity among the vocabulariesaccording to the semantic vector corresponding to each vocabulary so as to determine a semantic relationship among the vocabularies. According to the method, the semantic relation of the power grid dispatching major can be quickly and accurately extracted by means of the neural network model, the subjective influence of dispatching personnel is avoided, and the work load of the dispatching personnel is reduced.

The invention discloses a training method and device for a multi-language semantic representation model, electronic equipment and a storage medium, and relates to the field of natural language processing based on artificial intelligence. According to the specific implementation scheme, the method comprises the steps of: adopting a plurality of training corpora containing multiple languages to train a multi-language semantic representation model, so that the multi-language semantic representation model learns the semantic representation capacity of the various languages; for each training corpus in the plurality of training corpuses, generating a corresponding hybrid language corpus, the hybrid language corpus comprising corpuses of at least two languages; and training the multi-language semantic representation model by adopting each hybrid language corpus and the corresponding training corpus, so that the multi-language semantic representation model learns semantic alignment information of different languages. According to the technical scheme, the multi-language semantic representation model can learn semantic alignment information between different languages, semantic interactionbetween different languages can be achieved on the basis of the multi-language semantic representation model, and practicability is very high.

The invention discloses a multimedia-based multi-platform multi-terminal multi-language mutual translation method device, comprising a server end and a client end, the server end is a computer installed with a parsing program and a database, and the client end is a browser installed The terminal equipment of the translation software client program, the database is the sum of multimedia, language semantics and language voice information, multimedia is pictures, videos, animations and sounds, language semantics are words, words, sentences and articles, language voices are associated with semantics pronounce. The user sends an instruction to the server to translate the target language by operating the terminal device. After receiving the instruction, the parsing program on the server side translates the target language according to the corresponding relationship between semantics and pronunciation, and semantics and multimedia. The pronunciation and multimedia are sent to the client, and the client user receives the semantics, pronunciation and multimedia of the target language to achieve the purpose of translation. The invention is convenient, quick, intuitive and accurate, and vivid.

The invention relates to the technical field of speech recognition and discloses a speech recognition control method and device, wherein the speech recognition control method comprises: storing voiceprint characteristics of a first audio signal acquired to a local voiceprint characteristic database, acquiring voiceprint characteristics of a second audio signal, matching the voiceprint characteristics of the second audio signal with voiceprint characteristics of each first audio signal in the local voiceprint characteristic database, and if it is determined that the voiceprint characteristics of the second audio signal gain are fully matched with those of any first audio signal in the local voiceprint characteristic database, transmitting corresponding control commands to a controlled device. The speech recognition control method and device have the advantages that the problem is effectively avoided that since existing speech recognition devices recognize speech through a networked language semantic database, response is too slow in places with poor network signal or zero WiFi and user experience is poor; the problem is effectively avoided that incorrect speech of local accent leadsto inaccurate recognition; therefore, speech recognition rate is effectively increased.

The invention provides a network teaching real time voice analysis method, comprising steps of capturing a voice input, executing the real-time identification on the voice input, analyzing the identified voice input in order to mark the possible fault in the voice, and processing a text in order to extract a context dialogue prompt, wherein the voice input comprises the voice coming from the user or at least one another speaker. The real time recognition comprises a step of using automatic voice reorganization ASR to convert the voice input to the text. The context dialogue prompt is used for detecting at least one of candidate voice, a candidate work and a candidate word group in order to correct. The invention has advantages that the network teaching real time voice analysis method provides the real-time and passive monitoring to user voice, and does not need the active participation of the user. The method is highly interactive and highly, and can use the context and the language semantic.

The invention discloses a posture sensing-based body language semantic recognition method, a body language semantic recognition device and a storage medium. The method includes the following steps that: the target state information of arm joints and palm joints is acquired through posture sensing; state values corresponding to the target state information are searched in a first preset relationship mapping table, the state values of the arm joints and palm joints are combined, so that body language identification codes can be generated; and corresponding control commands are searched in a second preset relationship mapping table according to the body language identification codes, and operation on a target object is realized according to the control commands. According to the posture sensing-based body language semantic recognition method, the body language semantic recognition device and the storage medium of the invention, by means of the posture sensing, the state information of thearm joints and palm joints can be acquired more directly; the control commands corresponding to the state information are searched in the preset relationship mapping table according to the obtained state information; the operation on the target object is realized according to the control commands; and therefore, the efficiency of body language semantic recognition can be improved.

The invention relates to a multi-modal semantic fusion human-computer interaction system and method for virtual experiments. The system comprises an interactive information integration module, and further comprises an interactive information acquisition module, an interactive intention reasoning module and an interactive task straight-going module. The interactive information module adopts a multi-modal fusion model to accurately identify the real intention of an operator, and provides the acquired information for the interactive intention reasoning module; the interaction intention reasoningmodule is used for recognizing the interaction intention of the user according to the gesture semantics and the language semantics in combination with the current interaction scene of the scene and predicting potential interaction behaviors; and the interaction task execution module generates an experiment action expected by the user according to the interaction action predicted by the interactionintention reasoning module, generates a response experiment effect, returns corresponding operation feedback, and finally outputs the experiment effect and the feedback to the user through differentchannels. According to the invention, the problem of interaction difficulty in the current virtual experiment is solved.

The invention provides a shared handheld man-machine interaction device and a business mode thereof. An interactive platform, a central server, a cabinet and an interactive terminal are connected through a network. A family voice interaction connection is established for a user of the interaction terminal; by utilizing a service background AI system, a language semantic recognition system, biological feature recognition, virtual agent and an actuator system are constructed; voice semantic recognition and interactive voice response are completed; machine learning and decision management are perfected; interactive voice response questions of weather, traffic, clock, arithmetic, calendar, science, encyclopedic knowledge learning and the like are provided to a user; meanwhile, the real-time positioning data uploaded by the interaction terminal and the interaction platform can be received through the central server, and the data is compared with a set user age group, man-machine interactioncontent options and safety options of a central storage to judge whether to trigger an alarm or not; the handheld man-machine interaction device is light and handy and is convenient to carry; the hand-held use can be realized when people go out for activities; indoor voice control can be realized; the multifunctional voice interaction backpack can be put into a backpack, a luggage case, a file box and the like for monitoring important articles, can provide personal and property safety monitoring and alarm services for users, and can provide voice interaction responses in aspects of ancient poems, nursery rhymes, children's vocabularies, music, idioms, stories, English dialogue translation, life encyclopedia information, food recommendation and the like for the users.

The invention relates to a sentence meaning composition relation lamination identification method based on central blocks, and belongs to the technical field of computer science and Chinese information processing. The sentence meaning composition relation lamination identification method based on the central blocks is based on contemporary Chinese language semantics, and solves the problem of Chinese sentence meaning composition relation identification in a Chinese sentence meaning structural model. The sentence meaning composition relation lamination identification method based on the central blocks gives a concept of 'layer central block' firstly, and achieves effective mapping from a syntactic structure to a sentence meaning structure. A sentence meaning composition relations identification problem is divided into three layer relation identification problems, namely a relation between predicates, a relation between basic case and the predicates, and a relation between general case and sentence meaning compositions. Central block identification algorithm, basic case identification algorithm, general case identification algorithm, identification algorithm of the relation between the predicates, identification algorithm of the relation between the basic case and the predicates, identification algorithm of the relation between the general case and the sentence meaning compositions are respectively provided. A computer can analyze and acquire the sentence meaning composition relation in high accuracy rate and efficiency, and study of the Chinese sentence meaning structural model is further promoted.

The invention relates to the technical field of computers, and discloses a semantic checking method and system for multi-language mixed text. The method comprises the following steps: S1, identifying a main language and an auxiliary language, and extracting a translation auxiliary language; S2, conducting MLM model prediction; S3, preferably replacing translation candidate results; and S4, performing semantic checking on the main language. By designing an MLM model prediction module and a translation candidate result optimization module and using part-of-speech tagging screening and word vector similarity comparison technologies, a result really conforming to the current context can be accurately screened from a plurality of translation results of polysemy words, correct expression of the translation results to original meaning is ensured, and the accuracy of the translation results is improved; and therefore, reliable input is provided for a semantic checking model.

The invention provides a natural language understanding method in a small sample scene; the method proposes pre-training model language semantic representation, intention recognition and slot position recognition, and introduces tag semantics; then a linear space mapping method is used to draw a semantic representation distance, a gating network is established, slot information and intention information are fused, and abstract label transition probability is used to achieve the purpose of rapid learning and understanding in different fields. According to the method, the intention of the problem can be better judged in a small sample scene, and the slot position of the problem is identified, so that the problems of insufficient data, too high data labeling cost and too high model migration cost under a natural language understanding task of a task type dialogue system are well solved.

A computer-implemented method is provided. The method comprises: receiving, from a text field of a user interface, input text strings containing medical data; identifying, based on language semantics and grammatical structure, a keyword and one or more data values from the input text strings; providing the keyword as an input to a query of a medical data category database to obtain one or more categories associated with the keyword, the one or more categories including a first category; retrieving, from a template database, a first template associated with the first category, the first template including one or more fixed blocks and one or more variable blocks; inserting a data value corresponding to the keyword into a variable block of the first template to generate a replacement text string; and displaying the replacement text string in place of the input text strings in the text field.

The invention provides a method and device for automatically extracting a bilingual dictionary from pictures. The method for automatically extracting the bilingual dictionary from the pictures includes the following steps that a first language picture semantic annotation device and a second language picture semantic annotation device are used for annotating at least two pictures respectively; according to the output of the first language picture semantic annotation device and the output of the second language picture semantic annotation device, at least two bilingual sentence pairs both of which comprise first language semantic labels and second language semantic labels are generated; the EM algorithm is adopted for carrying out word aligning on the generated bilingual sentence pairs and entries of the bilingual dictionary are extracted. According to the method and device for automatically extracting the bilingual dictionary from the pictures, at least two bilingual sentence pairs are generated, the EM algorithm is adopted for carrying out word aligning on the bilingual sentence pairs, and therefore the entries of the bilingual dictionary can be automatically extracted from the pictures. Thus, dependence on a bilingual text and webpage resources is not needed and the method for automatically extracting the bilingual dictionary from the pictures is easy to achieve and high in accuracy.

The invention relates to a natural language semantic extraction method and system. The method comprises the steps of: performing word segmentation on a target file with sentences as units to obtain word segmentation units; analyzing whether a plurality of word segmentation units in one sentence form a keyword or not; in response to one or more keywords formed by a plurality of word segmentation units in one sentence, extracting the sentence containing the keywords; performing syntactic analysis on sentences containing the keywords to obtain a syntax tree; and extracting effective keywords and corresponding semantic tags from the syntax tree. According to the method and the system provided by the invention, the keyword meeting the information recommendation requirement is extracted from the content of the related target file, and the effective keyword is obtained by performing grammatical analysis on the sentence containing the keyword and splitting and reconstructing the syntax tree, so that the real intention of the target file can be obtained, and the understanding degree of the target demand of the user is improved.

In one example, a method of extracting patient information for a medical application comprises: receiving patient data of a patient; processing the patient data using a learning system with Artificial Intelligence (AI)-assisted clinical extraction tool, the processing comprising: extracting, based on a trained language extraction model that reflects language semantics and a user's prior habit of entering other patient data, data elements from the patient data and data categories represented by the data elements, and mapping at least some of the extracted data elements to pre-determined data representations based on the data categories; populating fields of a data record of the patient based on the pre-determined data representations; and storing the populated data record in a database accessible by the medical application.

The embodiment of the invention relates to the field of artificial intelligence, and discloses a sign language recognition method and device based on a neural network model, equipment and a medium. The method comprises the following steps of: obtaining to-be-recognized sign language information input by a user and collected by a shooting device, the sign language information comprising conversion mode sign language information and the to-be-recognized sign language information; inputting the sign language information into a pre-trained sign language recognition model to obtain first sign language semantic text information corresponding to the conversion mode sign language information and second sign language semantic text information corresponding to the to-be-recognized sign language information; determining a target language mode corresponding to the conversion mode sign language information according to the first sign language semantic text information; and converting the second sign language semantic text information into one or more of voice, characters and sign languages corresponding to the target language mode. Efficient and intelligent sign language recognition is achieved, and the sign language recognition flexibility is improved. The invention relates to a block chain technology, for example, sign language information can be written into a block chain so as to be used for scenes such as data evidence obtaining.