36 results about "Function word" patented technology

Techniques are provided for determining collaborative notes and automatically recognizing speech, handwriting and other type of information. Domain and optional actor/speaker information associated with the support information is determined. An initial automatic speech recognition model is determined based on the domain and/or actor information. The domain and/or actor/speaker language model is used to recognize text in the speech information associated with the support information. Presentation support information such as slides, speaker notes and the like are determined. The semantic overlap between the support information and the salient non-function words in the recognized text and collaborative user feedback information are used to determine relevancy scores for the recognized text. Grammaticality, well formedness, self referential integrity and other features are used to determine correctness scores. Suggested collaborative notes are displayed in the user interface based on the salient non-function words. User actions in the user interface determine feedback signals. Recognition models such as automatic speech recognition, handwriting recognition are determined based on the feedback signals and the correctness and relevance scores.

The invention relates to an automatic extraction method for text labels in combination with theme model and semantic analyses, pertaining to the technical field of computer application. The method comprises pre-treatment, LDA modeling, context analyses and label extraction.The pre-treatment comprises following steps: removing low-frequency words, removing stop words and removing label information, wherein stop words are auxiliary words without any information, words showing sentence grammar structures, all function words and punctuations. The LDA modeling process comprises following steps: obtaining two matrixes after processing the LDA model: one is a file-theme matrix of N*K with each element corresponding to a hidden theme distribution of each file and the other is a K*M theme-word matrix with each element corresponding to a word distribution of each theme. Based on a conventional counting method, the method takes correlations of words in files into consideration and fully utilizes one key feature of context information so that label information of files is obtained.

The present invention is directed to a system for managing business knowledge expressed as statements, preferably sentences using a vocabulary, where such statements may be automated by the generation of programming language source code or computer program instructions. As such, the present invention also manages software design specifications that define, describe, or constrain the programming code it generates or programs with which it or the code it generates is to integrate. All information managed within the present invention is maintained within a relational database that is encapsulated within an object-oriented model. Each object in this model is subject to version control and administration using permissions. Each user of the system is an object and belongs to one or more groups. Users and groups may be granted privileges. Objects may be created, examined, used, modified, deleted, or otherwise operated upon only if corresponding permission or privilege has been granted. The vocabulary managed by the present invention consists of the function words commonly used in a language, such as the auxiliary verbs, prepositions, articles, conjunctions, and other essentially closed parts of speech in English, as well as open parts of speech, such as nouns, verbs, adjectives, and adverbs.

The invention discloses a computer text input method and a system together with a maintenance method and a maintenance device of the thesaurus. The method includes the following steps: pre-storing a deficiency thesaurus of function words; storing the text information input through the computer text input system in the user thesaurus and count the word input frequency; searching out whether a same word as the function word in the thesaurus of function words exists in the user thesaurus and delete the same word from the user thesaurus; analyzing the word frequency of the user thesaurus and merge the text meeting the special requirement of the matching frequency larger than one. The invention can reduce the user thesaurus occupation of stored resources and computing resources and improve input efficiency and accuracy through the maintenance of user thesaurus. The invention also can select candidate word from the maintained user thesaurus for input choice according to the word frequency, thus further improving the input efficiency and accuracy.

The invention relates to a Chinese author identification method based on a double-layer classification model and a device for realizing the Chinese author identification method, belonging to the field of information security. Aiming at the problem of low identification accuracy caused by excessive authors, an author grouping layer is added in an author identification model; each author is represented into an author vector; authors are grouped by a clustering algorithm; a second layer is an author identification layer; a dependence relationship, a function word, a punctuation mark and a word class mark are extracted from the second layer to use as characteristics; and author identification is carried out in the group. According to the method or the device, the problem that the identification accuracy is lowered because of excessive authors can be effectively solved. Meanwhile, with a proposed characteristic dimensionality reduction and optimization method based on a main ingredient analysis method, the problem that the identification accuracy is affected by noise comprised by a high-dimensionality characteristic vector is solved. The Chinese author identification method can be applied to the author textual research field of a literature and also can be applied to the field of information security, such as copyright protection.

The invention relates to an English-Burmese bilingual parallel sentence pair extraction method and device based on BiLSTM-CNN, and belongs to the technical field of natural language processing. The method comprises the following steps: firstly, pre-training a bilingual word vector through a Muse tool; secondly, performing function marking on the sentence by utilizing the characteristics of the Burmese virtual words and the Burmese assistant words for identifying the subject-called guest of the Burmese, splicing syntactic structure information of each word into a word vector, encoding the sentence by using BiLSTM-CNN, and taking an output probability as a condition for measuring whether the sentence is a parallel sentence pair or not. According to the above steps, the BiLSTM-CNN-based British-Burmese bilingual parallel sentence pair extraction device is prepared through functional modularization. Compared with a traditional bilingual parallel sentence pair recognition system, the methodand the device are simpler. Experimental results show that the method and the device are superior to a baseline system in the aspects of accuracy, recall rate and other indexes. The accuracy is generally improved.

The invention relates to a machine language information processing technology. For the purpose that the machine imitates logic thinking method of human body to understand language and master grammar function, a presentation can be made from sentence structure of a subject, a predicate, an object, an attribute, an adverbial modifier and a complement to theory and application of a noun, a verb, an adjective, a quantifier, an adverb and a function word, and the analysis process of the function of each part can be demonstrated to be used as language teaching demonstration and provide basic exercise for language learning. The method provides each language with a grammar function of analyzing, judging and understanding language information, the grammar function is established on a commonly used and communicating platform, so that the machine can not only recognize language information, but also apply the language information to inter-translate and exchange between languages.

The invention discloses a clue management method and device, a terminal and a computer readable storage medium. The clue management method comprises the steps of obtaining an original customer name corresponding to a target customer; Identifying a location word group in the original customer name; Judging whether continuous fields in the original customer name are the same as brand names in a brand word bank or not; Taking a continuous field which is the same as the brand name of the brand word bank in the original customer name as a brand keyword; If no continuous field in the original customer name is the same as the brand name in the brand word bank, taking the field between the place word group and the enterprise function word group as a brand keyword; Judging whether the name of a cooperative customer is consistent with a place phrase and a brand keyword or not; And if yes, marking the target client as a cooperative client. According to the technical scheme, the clue library is subjected to data analysis, and the client, coinciding with the cooperative client, in the target client can be marked as the cooperative client, so that the business personnel can be prevented from repeatedly following the client.

The invention provides a method for establishing a user knowledge demand model by utilizing a Gaussian mixed model for the first time. Firstly, high-dimensional vectors of function words are generated by considering the semantic information of the function words based on a skip-gram model of knowledge base training word2vec, then the Gaussian mixed model is trained by utilizing selected knowledge corpus set, multiple Gaussian distributions are applied to describe the probability distributions of function word knowledge demands of a user, an EM method is applied to optimize parameters of the Gaussian mixed model; finally, the mapping relation between the words and entries is established, a knowledge entry demand model of the user is obtained, and knowledge entries, most possibly interested by the user, in a knowledge base are calculated on the basis and are pushed to the user. The established Gaussian mixed model can more closely fit the user knowledge demand model, and the knowledge push accuracy rate is improved.

The invention relates to a document classifying method based on a network measure index. The document classifying method comprises a sample training phase and a document classifying phase. The sample training phase comprises the first step of sample collecting, the second step of text segmenting, the third step of word class analyzing, the fourth step of function word and name removing, the fifth step of word frequency counting, the sixth step of characteristic set Vd establishing, the seventh step of characteristic network peak establishing, the eighth step of characteristic network edge establishing, the ninth step of average degree calculating, the tenth step of cluster coefficient calculating, the eleventh step of characteristic path length calculating and the twelfth step of network measure index interval obtaining. The document classifying phase comprises the first step of processing a document to be classified and the second step of judging document classification. According to the document classifying method, classifying is accurate, classifying efficiency is high, the problem that according to an existing classifying method, scientific and technical literature, novels and prose cannot be distinguished is solved, and a scientific classification method and a theoretical foundation is laid for automatic distinguishing of the scientific and technical literature, the novels and the prose.

The invention discloses a dialogue generation method and device based on two-stage decoding, a medium and computing equipment, and the method comprises the steps of dividing a dialogue reply generation process into two decoding stages, firstly inputting a dialogue context into a dialogue generation model, and mapping the dialogue context into a word embedding vector; inputting a word vector into a context self-attention encoder to obtain a feature vector of a dialogue context, inputting the feature vector into a first-stage Transformer decoder, and decoding to generate a notional word sequence; inputting the notional word sequence into a notional word sequence encoder to obtain a feature vector of the notional word sequence; and finally, inputting the context and the feature vector of the notional word sequence into a second-stage Transformer decoder, and decoding to generate a final reply. Through the two-stage decoding process, interference of the virtual words which are high in frequency but lack semantic information on the notional words is prevented, and therefore reply relevance and information amount are improved.

The embodiment of the invention discloses an information recommendation method and device and electronic equipment. The method comprises the following steps of: firstly, obtaining a keyword input in aretrieval interface; extracting to-be-retrieved functional words capable of representing categories or features of POIs from the keywords; then, according to the corresponding relationship between the function words and POIs, obtaining a point of interest POI corresponding to the to-be-retrieved functional word. In order to achieve accurate recommendation of information, POIs (Point Of Interest)extracted according to the to-be-retrieved functional words are set as candidate POIs; then the association degree between the function words to be retrieved and each candidate POI is calculated, andfinally the candidate POIs are selected and recommended according to the association degree, so that the technical problem that in the prior art, the fitness degree of POI recommendation and user retrieval intention is not high is solved, and the POI recommendation accuracy is improved.

A foreign language study book and a replaying device using the book are provided. In the foreign language study book, learning target foreign language words are distinguished as different words based not on the spelling or pronunciation unit but on the meaning unit of each word, and the words having different meanings are arranged by being dispersed over the whole inclusion area of the study book. Since grammatical information such as matching word information inherent to words, word order information, and function word information as well as the spelling and pronunciation can be learned synthetically, immigrants can develop foreign language vocabulary and composition ability as a mother tongue in a short time.

The invention discloses a method for semantic recognition and graph recommendation. The method comprises the steps that words and symbols input by a user in an input method are used to judge the integrity degree of the user's contents at first, and then function word parts in the contents will be filtered and notional word parts will be reserved according to word natures of the input words in the contents; a cloud side service is utilized synchronously, the filtered notional words are transmitted to the cloud side service according to the language input by the user, the cloud side service recommends hot graphic documents selected by users recently according to meanings of the notional words, the documents are presented on an input method interface for selection of the user, and the document will be automatically sent to an instant chat window after the user selects the graphic document; and operations are simple, execution efficiency is high, and the user can search for a graph without leaving a current application.

The embodiment of the invention discloses a man-machine interaction intention analysis method and device, computer equipment and a storage medium. The method comprises: picking up semantic interaction voice and converting the semantic interaction voice into a semantic text; performing syntactic dependency analysis to obtain an analysis result; judging whether punctuation marks exist in the analysis result or not; if the first punctuation mark exists, cutting off the analysis result according to the position of the first punctuation mark to obtain two clauses; determining a core relationship and a relationship between the advertent of the clause where the core relationship is located and the head word; determining whether the semantic text contains effective information or not; if not, judging whether the end of the semantic text is a virtual word or not; if yes, deleting the virtual words at the end of the semantic text; if not, retrieving a core relationship; and judging whether the semantic text contains effective information or not by combining the subscript length of the core relationship. By implementing the method provided by the embodiment of the invention, the problem that the current semantic service cannot accurately judge the real intention of the user is solved, so that the semantic service understanding is more accurate.

The invention discloses a method and a device for enhancing grammar error correction data based on a real error pattern. The method comprises the following steps: acquiring a to-be-noise-added statement and a noise adding strategy set; determining the noise adding probability of each word in the statement to be subjected to noise adding; randomly selecting a noise adding strategy from a noise adding strategy set according to the noise adding probability to carry out noise adding processing on the to-be-noise-added word; and constructing parallel statement pairs according to the error statements subjected to noise addition processing and the correct statements before noise addition processing. The noise adding strategy set comprises a real error pattern-based replacement strategy, a synonym replacement strategy, a function word replacement strategy, a similar spelling replacement strategy and a flexion replacement strategy. According to the embodiment of the invention, through introduction of real errors and simulation of various real errors, high-quality artificial error enhancement data which is more real and closer to real errors of learners can be generated; and various grammar errors can be manufactured through various types of noise schemes, and the method and the device can be widely applied to the technical field of data processing.

The embodiment of the invention discloses a part-of-speech tagging method, device and equipment and a storage medium. The method comprises the steps of obtaining an original statement; taking the original statement as the input of a part-of-speech tagging model to obtain the part-of-speech of each word in the original statement, wherein the part-of-speech tagging model is obtained by training a neural network model based on a virtual word part-of-speech corpus and a general word part-of-speech corpus. Through the technical scheme provided by the embodiment of the invention, the part-of-speechtagging accuracy of the words in the original statement is improved.

The embodiment of the invention provides a pronunciation dictionary generation method, a word speech recognition method, a word speech recognition device, electronic equipment and a storage medium. The pronunciation dictionary generation method comprises the steps: acquiring a training corpus which comprises a first phoneme sequence corresponding to one or more notional words, and a pronunciationrule corresponding to the language to which the notional word belongs; constructing one or more function words according to the pronunciation rule, wherein the function words have corresponding a second phoneme sequence; and generating a pronunciation dictionary by adopting the notional words, the first phoneme sequence, the function words and the second phoneme sequence. According to the method,the data volume of the pronunciation dictionary is ensured, and a pronunciation dictionary with sufficient words can be generated by using training corpora less than that for training a common pronunciation dictionary when facing unknown small languages, so that pronunciation of the to-be-identified words is accurately identified by increasing little corpora with large corpora.