267 results about "Semantic relevance" patented technology

An authoring method includes generating an authoring interface configured for assisting a user to author a text string in a source language for translation to a target string in a target language. Initial source text entered by the user is received through the authoring interface. Source phrases are selected that each include at least one token of the initial source text as a prefix and at least one other token as a suffix. The source phrase selection is based on a translatability score and optionally on fluency and semantic relatedness scores. A set of candidate phrases is proposed for display on the authoring interface, each of the candidate phases being the suffix of a respective one of the selected source phrases. The user may select one of the candidate phrases, which is appended to the source text following its corresponding prefix, or may enter alternative text. The process may be repeated until the user is satisfied with the source text and the SMT model can then be used for its translation.

There is disclosed a system and method for using a knowledge representation to provide relevant information based on environmental inputs. In an embodiment, the system and method considers environmental information from members in a crowd to generate a pool of interests based on the semantic relevance concepts associated with those interests. The most prominent concepts of interest may then be the basis for presenting content to the crowd as a whole. In another embodiment, environmental inputs and other surrounding inputs are considered as a user context. The concepts may be identified as relevant from the environmental context and used to present information relevant to the user given his or her surroundings.

The technology disclosed provides a so-called “joint many-task neural network model” to solve a variety of increasingly complex natural language processing (NLP) tasks using growing depth of layers in a single end-to-end model. The model is successively trained by considering linguistic hierarchies, directly connecting word representations to all model layers, explicitly using predictions in lower tasks, and applying a so-called “successive regularization” technique to prevent catastrophic forgetting. Three examples of lower level model layers are part-of-speech (POS) tagging layer, chunking layer, and dependency parsing layer. Two examples of higher level model layers are semantic relatedness layer and textual entailment layer. The model achieves the state-of-the-art results on chunking, dependency parsing, semantic relatedness and textual entailment.

The invention discloses a personalized information push system. Including: logo for the current web page that the theme of a number of topics Term Extraction Module website theme for the treatment of pushing all of the business users gave information in accordance with the semantics of the relevant calculation will be calculated results related to the semantic sort of calculation module. As well as the semantic relevance to sort the results will be consistent with the conditions of push personalized information presented to users of business information Push module. The invention also discloses a method push personalized information. The invention can attribute depending on the user's static and dynamic preferences, personalized information services push.

A method, apparatus, and medium for generating a grammar network for speech recognition and a dialogue speech recognition are provided. A method, apparatus, and medium for employing the same are provided. The apparatus for generating a grammar network for speech recognition includes: a dialogue history storage unit storing a dialogue history between a system and a user; a semantic map formed by clustering words forming each dialogue sentence included in a dialogue sentence corpus depending on semantic correlation, and generating a first candidate group formed of a plurality of words having the semantic correlation extracted for each word forming a dialogue sentence provided from the dialogue history storage unit; a sound map formed by clustering words forming each dialogue sentence included in the dialogue sentence corpus depending on acoustic similarity, and generating a second candidate group formed of a plurality of words having an acoustic similarity extracted for each word forming the dialogue sentence provided from the dialogue history storage unit and each word of the first candidate group; and a grammar network construction unit constructing a grammar network by combining the first candidate group and the second candidate group.

A microprocessor executable method and system for determining the semantic relatedness and meaning between at least two natural language sources is described in a prescribed context. Portions of natural languages are vectorized and mathematically processed to express relatedness as a calculated metric. The metric is associable to the natural language sources to graphically present the level of relatedness between at least two natural language sources. The metric may be re-determined with algorithms designed to compare the natural language sources with a knowledge data bank so the calculated metric can be ascertained with a higher level of certainty.

The invention provides a method for constructing correlation networks of keywords of natural language texts. The method includes steps of constructing dictionaries of the keywords and segmenting words of target corpuses according to the dictionaries to obtain a plurality of words; performing statistics on front and rear word correlation frequencies of the multiple obtained words on the basis of N-element statistic language models; training the language models by the aid of neural networks under training conditions which are the frequencies obtained by means of statistics, and acquiring word vectors; computing the similarity degrees of the word vectors of each two corresponding words and generating semantic correlation between each two corresponding words; generating the correlation networks of the keywords of the texts according to the level of the semantic correlation between the corresponding words. The similarity degrees of the word vectors of each two corresponding words are used as measurement for the semantic correlation of the two words. The method has the advantage that the accuracy of the correlation networks of the texts in relevant items can be effectively improved by the aid of the method.

A method of providing semantically relevant, ranked search results is described. An assignment of an occurrence of a term within a document to an ontological node is received based on usage of the occurrence of the term within the document. An accuracy value associated with the assignment is determined by comparing the assignment with a baseline assignment. A consistency value associated with an assignor is determined based on a repeatability of a plurality of assignments by the assignor. An assignor relevance score for the occurrence of the term is determined by statistically combining the accuracy value and the consistency value. An overall semantic relevance value for the occurrence of the term is determined by statistically combining the assignor relevance score determined for a plurality of assignors. The documents are ranked based on the overall semantic relevance value and presented in response to a search request.

Embodiments of the invention provide a text relevance calculating method and a device thereof. The method comprises the following steps: receiving a first character string and a second character string; calculating a text relevance characteristic value of the first character string and the second character string, and calculating a semantic relevance characteristic value of the first character string and the second character string; fitting the text relevance characteristic value and the semantic relevance characteristic value into a relevance characteristic value of the first character string and the second character string based on the logistic regression model. The text relevance calculating method and the device thereof increase the precision of relevance judgment, save storage space and reduce the cost.

The invention discloses an engineering drawing material information extraction method based on a template, comprising the following steps of: generating a table figure, words and filling rule description information of table units by using figure software to generate a table extraction template; reading and identifying basic figure element type information, figure property parameter information, rule description information and topological structure information which are contained in the extraction template; analyzing the feature of the extraction template to form table feature description according to the topological structure information; circularly reading and identifying basic figure element types and figure property parameter information in a CAD (Computer Aided Design) design drawing and then identifying table frames according to table features to form table frame integrations; circularly identifying the element of each table frame integration and then reading and identifying the basic figure element types and the figure property parameter information; and extracting material information and storing the material information in a database. The invention improves the extraction precision of the table features and ensures the extracted semantic relevance and the extraction accuracy of the material information.

The invention discloses a searching method and searching equipment. The searching method comprises the following steps: receiving a search keyword input by a user, acquiring at least one search document relating to the search keyword based on the search keyword, and determining the semantic vector of the search keyword; according to the semantic vector of the search keyword and the semantic vector of each search document, calculating the semantic correlation between each search document and the search keyword; according to the semantic correlation, sequencing the search documents and outputting the search documents to the user according to the sequencing result. The semantic vector of the document of a search engine is acquired according to the semantic vector of a document tile, the semantic vector of a document identifier and the semantic vector of a document category, so that the semantic information of the document can be expressed accurately, the precision of the semantic correlation between the document acquired by calculation and the search keyword is improved, the search precision of the search engine is improved, and the user experience on the search engine is improved.

The invention discloses a method for generating suggested keywords of sponsored search advertisement based on user feedback, which develops semantic relatedness between words and establishes a learning machine to evaluate the relatedness of word pairs by utilizing the feedback information of users on the relatedness of a small quantity of word pairs so as to select words which has high relatednesswith seed keywords describing advertiser products or service concepts as the suggested keywords to be recommended to advertiser users, wherein an active learning method is used for selecting a word pair sample with maximal information quantity to evaluate user relatedness so as to enhance the efficiency and the precision of generating related suggested keywords. The invention can effectively generate hundreds of to thousands of related words to be used as the suggested keywords to be recommended to the advertiser users corresponding to each seed keyword; in addition, the advertiser users carry out sponsored search bidding for the suggested keywords, thus the related click rate of the sponsored advertisement can be effectively enhanced, and the advertisement benefit is increased.

A method for network resource classification and identifying user interests based on the classification. The method uses a provided hierarchy of categories for classifying network resources, wherein each category is assigned a text item describing the category and the method includes obtaining resource description data collections corresponding to the network resources, and generating, using a semantic correlation algorithm, a category score vector of a network resource by comparing the resource description data collection to the text item assigned to each category in the hierarchy of categories, wherein the category score vector comprises a category score for each category in the hierarchy of categories, wherein the category score is determined based on at least a semantic correlation measure between the resource description data collection and the text item assigned to a corresponding category, wherein the plurality of network resources are classified based at least on the category score.

The invention relates to the field of automatic abstracting, and discloses a graph model-based automatic abstracting method. According to the technical scheme, an LDA probability topic model is applied to measurement of semantic correlation between sentences and improvement of the measurement effect of sentence correlation; and an idea of topic correlation and position sensitivity of the sentences is provided, so that abstract generation is relatively reasonable and effective. The method comprises the following steps: firstly, obtaining topic probability distribution of a document and word probability distribution of the topic through training the LDA topic model, determining the topic probability distribution of the sentences and effectively converting a semantic similarity measurement between the sentences into a similarity measurement problem of the topic probability distribution of the sentences; with the sentences as nodes, building edges by referring tothe cosine similarity and according to the semantic similarity between the sentences and generating a text graph representing the document; calculating the topic correlation between the sentences according to the topic probability distribution of the sentences and the topic probability distribution of the document; and calculating the position sensitivity and the like of the sentences according to the positions of the sentences in the document.

The invention discloses a cross-modal retrieval method based on a cyclic generation antagonistic network. The method designs a novel dual-channel cyclic generation antagonistic neural network and establishes semantic correlation of cross-modal data by training the neural network. Given different modal data can flow bi-directionally in the network, each modal data can generate another modal data through a group of generative antagonistic networks, and the generated data can be used as the input of the next group of generative antagonistic networks, so that the bi-directional loop generation ofdata can be realized, and the network continuously learns the semantic relationship among the cross-modal data. In order to improve the efficiency of retrieval, The method also uses threshold functionand approximation function to approximate the result of the middle layer of the generator to the corresponding binary hash code, and designs a variety of constraints to ensure the same mode, the similarity and cross-mode of the same kind of data, and the difference of data between classes, thus further improving the accuracy and stability of retrieval.

The invention relates to a computer assistant reporting and knowledge-base generation method including the following steps: a server system receives requests of information search from users and searches all structured and unstructured web pages and websites matching the search requirements of users on the internet and a third-party database and also feeds the search results back to a client digital terminal system after classification, de-emphasis and compilation while matching the third-party database; the server system records the search behavior of a user and detects the updating of information resource and catches and classifies the updated information in real time so as to remind the user of the updating of the information resource when the user goes on line; the user selects the required information from the search results and collects the information into the system and the knowledge mining method is adopted to generate reports and derive files. The invention has the beneficial effects that the efficiency of searching, collecting and classifying the information is enhanced for the user, and the unstructured information can be converted into structured information and the semantic relevance between information can be preserved by human-computer interaction.

The invention discloses a local spline embedding-based orthogonal semi-monitoring subspace image classification method. The method comprises the following steps of: 1) selecting n samples serving as training sets and the balance serving as testing sets from image data sets, wherein the training sets comprise marked data and unmarked data; 2) building an extra-class divergence matrix and an intra-class divergence matrix by using the marked data; (3) training data characteristic space distribution by using a whole and building a Laplacian matrix in a local spline embedding mode; 4) according to a local spline, embedding an orthogonal semi-monitoring subspace model, and searching a projection matrix to perform dimensionality reduction on the original high dimension characteristic; 5) building a classifier for the training samples after the dimensionality reduction by using a support vector machine; and 6) performing the dimensionality reduction on the testing sets by using the projection matrix and classifying the testing sets after the dimensionality reduction by using the classifier. In the method, the information, such as image sample marking, characteristic space distribution and the like, is fully utilized; potential semantic relevance among image data can be found out; and image semantics can be analyzed and expressed better.

The invention relates to a text abstract model generation algorithm for fusing information selection and semantic association. Firstly, based on the Decoder model, the attention mechanism is combinedto obtain enough information of input sequence. Secondly, a copy mechanism and a coverage mechanism are used to solve the problem of non-login words and clause repetition in summary generation. Then aselective network is designed to encode the original text twice to filter the redundant information. Finally, by comparing the semantic relevance of the original text and the abstract, the semantic relevance of the abstract can be corrected, and the semantic relevance of the abstract and the original text can be enhanced.

Ranking search results, comprises receiving a query string; retrieving a plurality of search results that include a corresponding plurality of target strings that relate to the query string; segmenting the query string and each of the plurality of target strings; pairing segments in the query string with respective segments in the target strings to form a plurality of combinations; retrieving a plurality of weights that correspond to the plurality of combinations based on a mapping of word combinations and their respective weights, wherein a weight measures semantic correlation between words in a word combination; and determining a weighted word length based on the weights corresponding to each of the plurality of target strings; and ranking the plurality of target strings based on their respective weighted word lengths. Alternatively, ranking search results includes determining a minimum weight of each inserted word with respect to segmented words in the query string; determining a minimum weight of each deleted word with respect to segmented words in the target strings; determining a total edit distance based at least in part on the minimum weight of each inserted word and the minimum weight of each deleted word; and ranking the target strings based on the total edit distances.

The invention relates to a network image retrieval method based on semantic analysis, which is used for extracting low-level features. Content-based image retrieval is performed on each type of feature to find out a visually-similar network image set. The related text information is used for semantic learning corresponding to each image in the network image set corresponding to each image in the network image set to obtain the semantic expression for the image query. The semantic consistency of the retrieval image set corresponding to various features on the text information is judged, the semantic consistency is used to measure the description capacities of various features, to endow the description capacities with different degree s of confidence. The semantics and semantic consistency of the image query are adopted to perform text-based image retrieval in the image base to obtain the semantic relevance of each image in the image base and the image query; the low-level features are adopted to perform content-based image retrieval on the image base to obtain the visual relevance of the each image in the image base and the image query; the semantics is fused with visual relevance through a linear function to ensure the image for the user to have both semantic and visual relevance.

The present invention uses an algorithm which evaluates learners' short free-text answers when the answer has as few as 10 words. The answer key uses only one correct answer, allowing instructors to ask learners to produce short open-ended text responses to questions. The algorithm automates the scoring of free-text answers, enabling instructors to embed such questions in online courses, and providing nearly immediate scoring and feedback on learners' responses. The algorithm is based on the semantic relatedness of the words in the learners' answer to the single correct answer. The semantic relatedness algorithm requires a dedicated domain specific index or collection of topic-focused documents (a corpus), which is created by an automated crawl mechanism that collects documents based upon descriptive domain keywords.

The invention relates to an image combined with deep learning and hashing methods. And a text cross-modal retrieval model. The invention provides a deep cross-modal Hash algorithm based on multilayersemantics in order to solve the limitation that a traditional cross-modal Hash method based on deep learning directly converts multi-label data into a single label problem when processing the multi-label data problem. And the similarity between the data is defined through a co-occurrence relation between the multi-label data, and the similarity is used as supervision information of network training. And designing a loss function which comprehensively considers multi-layer semantic similarity and binary similarity, and training the network, so that the feature extraction and Hash code learningprocesses are unified in one framework, and end-to-end learning is realized. According to the algorithm, semantic correlation information between the data is fully utilized, and the retrieval accuracyis improved.

A method of providing semantically relevant, ranked search results is described. An assignment of an occurrence of a term within a document to an ontological node is received based on usage of the occurrence of the term within the document. An accuracy value associated with the assignment is determined by comparing the assignment with a baseline assignment. A consistency value associated with an assignor is determined based on a repeatability of a plurality of assignments by the assignor. An assignor relevance score for the occurrence of the term is determined by statistically combining the accuracy value and the consistency value. An overall semantic relevance value for the occurrence of the term is determined by statistically combining the assignor relevance score determined for a plurality of assignors. The documents are ranked based on the overall semantic relevance value and presented in response to a search request.

A microprocessor executable method transforms unstructured natural language texts by way of a preprocessing pipeline into a structured data representation of the entities described in the original text. The structured data representation is conducive to further processing by machine methods. The transformation process is learned by a machine learned model trained to identify relevant text segments and disregard irrelevant text segments The resulting structured data representation is refined to more accurately represent the respective entities.

The embodiment of the invention discloses a model training method, device and equipment, and the method comprises the steps: obtaining a pre-trained dialogue generation model which comprises an encoder and a decoder; constructing a dialogue system evaluation model, wherein the dialogue system evaluation model takes questions and replies as input and takes scores corresponding to the replies as output; initializing parameters of an encoder in the dialogue system evaluation model according to the parameters of the encoder in the dialogue generation model; and according to the first training sample set, training the initialized dialogue system evaluation model to obtain a dialogue system evaluation model meeting a training ending condition, each training sample in the first training sample set comprising a question, a reply and a label score corresponding to the reply. The dialogue system evaluation model obtained by training through the method can evaluate the reply quality of the dialogue system from the perspective of semantic correlation, and the reliability of dialogue reply evaluation is improved.

The invention discloses a concept relation label drawing method based on a semantic co-occurrence model. The concept relation label drawing method includes obtaining co-occurrence context phrase of a concept relation pair through inverted searching with offset in a relation label drawing process based on the semantic co-occurrence model; obtaining a candidate relation label by combining shallow parsing and weighting word window filtering; and choosing a cluster label largest in weight after semantic relativity cluster is performed to serve as a concept relation label. By means of the concept relation label drawing method based on the semantic co-occurrence model, accuracy rate and recalling rate of concept relation drawing are improved, automatic drawing of few and scattered semantic relation concept relation labels among large number of concepts is achieved, and quality of concept label drawing is improved.

The invention relates to a fine-grained visual question-answering method combined with a multi-view attention mechanism. The guiding effect of specific semantics of the problem is fully considered. Amulti-view attention model is provided. A plurality of salient target areas related to a current task target (problem) can be effectively selected From multiple perspectives, region information related to answers is acquired in images and question texts, regional significance features are extracted in the images under the guidance of question semantics. The characteristic expression of finer granularity is realized; the multi-view attention model has the advantages that the multi-view attention model is constructed, the situation that a plurality of important semantic expression areas exist in the image is expressed, the depicting capacity is high, the effectiveness and comprehensiveness of the multi-view attention model are improved, and therefore the semantic relevance of image area significant features and question features is effectively enhanced, and the accuracy and comprehensiveness of semantic understanding of visual questions and answers are improved. The visual question-answering task is carried out by adopting the method, the steps are simple, the efficiency is high, the accuracy is high, the method can be completely used for business, and the market prospect is good.

The invention discloses a correlation-quality sequencing image retrieval method based on tag retrieval. The correlation-quality sequencing image retrieval method comprises the following steps of: firstly, automatically sequencing social images according to the correlation of images and tags, fusing the vision consistency among the images and the vision consistency of the semanteme correlation among the images and the tags by utilizing a normalization frame, and solving an optimization problem by utilizing an iteration algorithm so as to obtain a correlation sequence; and selecting the brightness, the contrast ratio and the color variety of the images to be used as quality characteristics of the images to balance previously returned images, overlaying correlation scores and quality scores by utilizing a linear model, and re-sequencing the images according to a final total score, and therefore, a correlation-quality sequencing algorithm is realized.

The embodiment of the invention provides a method and a device for predicting a knowledge graph. The method comprises the steps of acquiring a first entity and a second entity; establishing a first triangular sub-graph according to the first entity and the second entity; and determining a relation between the first entity and the second entity according to the first triangular sub-graph and an embedded model. According to the embodiment, the first entity and the second entity are acquired, the triangular sub-graph containing the first entity and the second entity is established, and the relation between the first entity and the second entity is determined according to the triangular sub-graph and the embedded model. In the embodiment, a relation type of the triangular sub-graph in the knowledge graph on the semantic level can be fully adopted, the relevance between object entities is acquired, and thus the problem of data sparseness caused by only consideration of direct semantic correlation of an entity pair is avoided.