66 results about "Natural language semantics" patented technology

A system, method, and computer program product for natural language understanding, including forming pairs of words and/or phrases in independent phrases in a sentence, wherein each word and/or phrase is paired with every other word and/or phrase; determining meanings for the words and/or phrases; comparing the meanings of each word and/or phrase with each meaning of the other word and/or phrase in the pair; selecting the pairs with the best relationships; combining highly-related pairs with other highly-related pairs to form longer groups of words; exchanging word meanings for the longer groups with word meanings having a weaker relationship to determine if the exchanged word meanings provide an overall stronger relationship; and forming longer and longer groups until the sentence is understood.

A relation detection model training solution. The relation detection model training solution mines freely available resources from the World Wide Web to train a relationship detection model for use during linguistic processing. The relation detection model training system searches the web for pairs of entities extracted from a knowledge graph that are connected by a specific relation. Performance is enhanced by clipping search snippets to extract patterns that connect the two entities in a dependency tree and refining the annotations of the relations according to other related entities in the knowledge graph. The relation detection model training solution scales to other domains and languages, pushing the burden from natural language semantic parsing to knowledge base population. The relation detection model training solution exhibits performance comparable to supervised solutions, which require design, collection, and manual labeling of natural language data.

The invention discloses a natural language semantic analysis system and method based on the depth neural network. The method comprises the steps that a knowledge map is built, a training set is inputted, and an N-Gram probability model is obtained, a matrix is obtained as an input by representing words as vectors using the word2vec, a deep belief network model is used for the entity identification and the input validation set, the classifier parameters and the input test set are adjusted, the group abilities of the models are tested, the knowledge graph method is adopted to apply reasoning to the entities in the descriptions of the language, and corresponding conclusions are obtained. Compared with the prior art, the natural language semantic analysis system and method based on the depth neural network uses the knowledge graph method to apply reasoning to the entities in the descriptions of the language and to obtain the corresponding conclusions, so that our natural language understanding abilities are provided not only with the capacity to understand the literal meaning, but also with logical reasoning and the understand of the meaning on a deep level, and the method has promotable and practical value.

The invention relates to a programming language capable of cross-language reusing, and discloses a method for programming and developing software application programs by using a group of certain brief term and syntax rule with natural language semanteme. Through the description of multiple programming language families and the display method of the same thought realizing parts in multiple language development of certain software with an abstract special function, the cross-platform, cross-compiling, convertible and embeddable unified development is realized. Through the compiling tool disclosed by the invention, the same set of codes programmed by the language can be compiled into the applications with different states under different platforms; and the platform characteristics are realized by corresponding compiling tools. Through the converting tool disclosed by the tool, the codes programmed by the language can be converted into other languages as C++, Java, C#, Pascal and Obj-C, and can also be converted into the codes directly executed by the ARM embedded chip. Through the standard developing library disclosed by the invention, the language can be directly embedded into other languages for mixed programming. Through the IDE and the simulator disclosed by the invention, the cross-platform debugging and simulation can be realized.

The invention relates to method and device for embedding and detecting a watermark in a document by using a computer system, in particular to method and device for loading a document watermark to an electronic text document by using a computer program so as to protect the electronic text document, belonging to the technical field of document protection. The computer program at least comprises the generation of the electronic text watermark, the loading and embedding of the watermark and the detection of the watermark. The method comprises the following steps of: firstly, generating watermark information from copyright information prepared to be added into the electronic document by using a watermark generating module; then, reasonably optimizing the analysis of a watermark carrier medium, the preprocessing of embedded information, the selection of an information embedding point, the design of an embedding mode and the control of embedding modulation by using a watermark loading and embedding module; and equivalently transforming set natural language semantics to embed the watermark through reasonably selecting a secret key. If a copyright dispute happens, a watermark detection module designed in the invention can be used for detecting and extracting the document watermark of the electronic document by using the preset secret key, and the result shows that the document watermark embedded in advance can provide an identification basis for solving the electronic document copyright dispute. Thus, the invention can be widely applied to document protection encryption, tracking, illegal distribution control and technical protection of copyrights.

The invention discloses a hierarchical semantic tree construction method and system for language understanding. The method mainly comprises the steps as follows: segmenting terms of a statement and loading a semantic knowledge base; recognizing all nodes of the statement according to an LV rule, and recognizing the level of the nodes according to semantic knowledge and term positions and collocations; generating a special node by punctuation at the end of the statement, and taking the special node as a root node of a semantic tree; merging the nodes according to generated node information, recognizing semantic side chunks of the statement, and taking a level-0 semantic side as a child node to be hung on the root node; circularly traversing all child nodes of the statement till no low-level semantic side exists, and taking the child nodes as leaf nodes to be hung on the child node. According to the hierarchical semantic tree construction method and system, under the condition that no syntactic resource exists, the semantic structure tree is obtained through semantic information and the term positions and collocations only, so that a computer can enter a deep semantic layer of a natural language, various processing of the natural language can be finished on the basis of understanding, the first step of semantic understanding of the natural language is realized, and the hierarchical semantic tree construction method and system can be applied to information retrieval, automatic abstraction, machine translation, text categorization, information filtration and the like.

The present invention discloses a natural language semantic understanding method and system. The method includes: obtaining text to be parsed; performing word segmentation and part-of-speech tagging on the text to be parsed; performing syntactic analysis on the text after word segmentation based on context-related information, and obtaining the Describe the natural word-dependent syntactic tree of each sentence in the text to be processed; construct a semantic tree according to the natural word-dependent syntactic tree and the pre-built ontology knowledge base; use the semantic tree to obtain semantic understanding results. The invention can improve the correctness and flexibility of natural language semantic understanding.

The invention discloses a digital cookbook generation method and system. The method comprises: firstly, creating a cooking term bank, wherein cooking terms have semantic descriptions, and the semantic descriptions are associated with corresponding cooking actions; secondly, writing a text cookbook close to a natural language semantic structure by cookbook editors through utilizing the cooking terms; and finally, compiling the text cookbook into a cookbook program by compiling software, and generating a digital cookbook with a standard data structure. According to the digital cookbook generation method and system, the learning of the cookbook editors to a computer programming language is weakened and simplified, the technical difficulty in writing cookbook programs by the cookbook editors is lowered, the key point of cookbook editing focuses on the digital cookbook content, the technical problems of existing digital cookbook editing are solved, the efficiency and effect of cookbook program writing are improved, and the security and reliability of the digital cookbook executed in a cooking device are ensured.

The invention discloses a conference summary generation method and device, a computer device and a storage medium. The method comprises the steps of obtaining sign-in video information of participantsstored in an attendance device; extracting a face image in the sign-in video information, and performing face recognition on the face image; extracting a voiceprint feature corresponding to each participant identifier in the sign-in video information from a preset voiceprint database; obtaining audio information and current time contained in the detection request; carrying out voiceprint featurerecognition on the audio information, determining a participant identifier corresponding to each piece of audio information, and carrying out voice recognition on the audio information; correlating the identifiers of the participants, the current time and the real-time voice text to obtain conference summary information; and performing semantic recognition on the conference summary information byadopting a natural language semantic recognition mode to obtain a target conference summary of the conference summary information. The conference summary generation method can improve the intelligentdegree of the conference summary and the generation efficiency of the conference summary.

Structured web pages are accessed and parsed to obtain implicit annotation for natural language understanding tasks. Search queries that hit these structured web pages are automatically mined for information that is used to semantically annotate the queries. The automatically annotated queries may be used for automatically building statistical unsupervised slot filling models without using a semantic annotation guideline. For example, tags that are located on a structured web page that are associated with the search query may be used to annotate the query. The mined search queries may be filtered to create a set of queries that is in a form of a natural language query and/or remove queries that are difficult to parse. A natural language model may be trained using the resulting mined queries. Some queries may be set aside for testing and the model may be adapted using in-domain sentences that are not annotated. The models may be tested using these implicitly annotated natural-language-like queries in an unsupervised fashion.

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 provides an artificial intelligence interaction method, and the method is used for solving the technical problem that as interaction information responses lack emotional characteristics, the interaction effect of an auxiliary artificial intelligence role is poor. The embodiment of the method comprises the steps that corpus resources of a role are obtained from a corpus and subjected to vectorization processing, and role corpus resources of the role are generated; natural language semantic recognition is conducted on the role corpus resources through a neural network model, a determined role is generated, sentence emotional characteristics of the determined role are generated, and personality characteristics of the determined role are generated. According to the artificial intelligence interaction method, combined with text words or emotion control information input by interaction information, by means of natural semantic recognition, an artificial intelligence assistant with personality is created, human participation is reduced on a large scale, the workload of development is relieved, and meanwhile the artificial intelligence assistant with a more natural emotional effect can be generated. The invention further provides an artificial intelligence interaction device.

The invention discloses an intelligent scoring method and device based on NLP, equipment and a storage medium. The method comprises the steps: acquiring a test question image to be scored; using an optical character recognition mode; identifying the test question image; obtaining identified content, dividing according to a preset statement division mode; performing statement segmentation on the identification content; obtaining each basic statement contained in the identification content; using an N-Gram model, performing content correction on each basic statement; obtaining a target text corresponding to each basic statement; finally, using a natural language semantic recognition mode, performing semantic recognition on each target text, to obtain the target semantics of each target text,and determine the comprehensive score corresponding to the test question image according to the target semantics. Character recognition is carried out through the test question image, test question review is carried out on the recognized characters in a natural language recognition mode, the scoring efficiency is improved, and the labor cost is saved.

The present invention relates to government service intelligent terminal equipment based on artificial intelligence and a use method thereof. A bar code, two-dimensional code identification, a contactless IC card reader, thermal paper printing, a touch display screen, a speaker, a noise reduction microphone, video collection equipment, a number keyboard, a license scanner and ticket printing equipment are included. The artificial intelligence technology is introduced, the speech recognition technology, the natural language semantic analysis technology, the contextual correlation analysis technology are used to process human-computer interaction, demand screening, process guidance, data collection and processing and other difficult issues, and a series of government affair matters such as consultation, inquiry and handling of government affair matters has been successfully realized in the intelligent terminal equipment.

The present invention relates to one kind of utilizing natural language sementeme to appoint objective rules and realize united coding of various information in natural language sementeme. The methodmakes the user capable of performing man-machine interaction in any familiar natural language and completing the coding of any natural languge text in united sementeme information. The united coding result with the sementeme information may be converted in a computer into various translation. The present invention can promote the development of various natural language processing technology.

The invention discloses an intelligent courseware generation method and device, computer equipment and a storage medium. The method comprises the following steps of: receiving a courseware generationrequest; obtaining basic field information contained in the courseware generation request in a web crawler mode; crawling knowledge point content containing the basic domain information as a base content; performing semantic analysis on the basic content in a natural language semantic recognition mode; obtaining semantic analysis results, classifying knowledge points in the basic content accordingto a semantic analysis result; obtaining core knowledge points and basic knowledge points; constructing a knowledge graph system corresponding to the basic domain information; the knowledge graph system corresponding to the basic domain information and the preset courseware template are used for generating the target electronic courseware, acquisition, analysis and classification of knowledge points containing the basic domain information are achieved, the corresponding knowledge graph system is obtained, then the target electronic courseware is generated, and the efficiency and the intelligent degree of electronic courseware generation are improved.

The invention discloses a template-based software defect automatic question and answer method, and belongs to the field of software maintenance, and the method comprises the steps: firstly, extractingan entity relationship triple from a bug corpus, and obtaining a natural language mode; Then determining an entity relationship in the triple; Acquiring a query template corresponding to the naturallanguage mode; Replacing an entity in the question q proposed by the user with an entity type to obtain a question q '; Comparing and searching the entity type in the q'with the entity type in the natural language mode, and calculating the similarity; Obtaining an SPARQL query mode of the question q according to the similarity and the entity in the question q; And finally, searching and executingthe SPARQL query mode of the question q on the entity relation graph constructed by the bug report, thereby obtaining the answer of the question q. According to the method, natural language semantic understanding is carried out through the template, the understanding effect is good, template matching is carried out based on the entity type, the search space is greatly reduced, and the efficiency of software defect automatic question answering is improved.

The invention discloses an intensive learning based medical natural language semantic network feedback extraction system and method. The system comprises a medical text big-data module, a medical ontology extraction module, a medical semantic network module and a quality medical semantic network database module; the medical text big-data module comprises an electronic medical record, biomedical literature, biomedical patents and network forum data; the medical ontology extraction module is used for extracting a series of biomedical entities from a medical text; the medical semantic network module takes the medical entities as nodes and form the network by taking ontological relations as connections; the quality medical semantic network database module forms high-quality structured data through the medical semantic network via crowdsourcing proofreading and experts proofreading and is used for intensive learning and training of the relationship extraction neural network.