37results about How to "Correct answer" patented technology

The question answering method and system of the invention relate to a method which realizes the questioning and answering by taking advantages of different websites to exchange information through network and a system for realizing the method. The method of the invention includes the following steps: presetting all the characteristics of the problem processed by cooperative websites in a question answering system; then the question answering system analyzes and forwards the question submitted by a user to the corresponding cooperation website which is able to solve the problem; the corresponding cooperation website displays the question in the web pages for the users browsing in the cooperation website and forwards answers of the question to the question answering system; finally the answers are forwarded to the corresponding questioning user by the question answering system. The invention makes full use of the wisdom of current website users by forwarding the questions of users to the cooperative website users, who browse the question and provide the answers, conducive for providing fast and accurate answers to the questioning users.

The invention discloses a method and a device for medical automatic question answering, a storage medium, and an electronic device, and belongs to the technical fields of natural language processing and computer artificial intelligence. Technical problems solved by the method and the device are how to use a natural language processing technology to select a matched answer in existing question answering results for a patient and recommend the answer to a user, so as to reduce workload of doctors and improve speed of question answering. Technical schemes comprise: 1, a method for medical automatic question answering; 2, a device for medical automatic question answering. The method comprises the following steps: S1, establishing a medical question answering knowledge base; S2, establishing aquestion answering model training dataset; S3, establishing a question answering model; S4, and training the question answering model and selecting an answer. The device comprises a medical question answering knowledge base establishing unit, a question answering model training dataset generating unit, a question answering model establishing unit, and a question answering model training and answerselecting unit.

The invention belongs to the technical field of computer intelligent conversation and provides an unstructured text-oriented intelligent question and answer method and system. The method comprises thesteps that S1, a coding layer performs coding on obtained text and question to obtain a text hidden vector and a question hidden vector; S2, an information fusion layer fuses the text hidden vector and the question hidden vector, and obtains a fused correlation vector group; and S3, a decoding layer performs decoding on the text according to the correlation vector group to obtain an answer of thequestion and outputs the answer. The answer is directly given for the raised question of the unstructured text, without establishing a question and answer library in advance; the type of the raised question is not limited; the returned answer is relatively accurate; and the data-driven design is adopted, so that big data is effectively utilized.

The embodiment of the present invention discloses a knowledge inference method applied to intelligent robot interaction, which comprises the following steps: acquiring a problem text and an inferencecorpus, wherein the inference corpus comprises unstructured information and triplet information; storing the unstructured information and the triplet information into a first memory module and a second memory module; calculating the correlation of the problem text with the unstructured information and the triplet information in the first memory module; performing weighted summation on the unstructured information and the triplet information in the second memory module and the correlation, and obtaining an output value; and generating an answer according to the output value. According to the aforementioned scheme, the scattered unstructured information and the structured triplet information are introduced and used as input information of the memory network to jointly infer the answer of thequestion, thereby giving a more accurate and satisfactory answer.

The invention discloses a seat-assisted question-answering method and system fusing semantic classification and a knowledge graph. The method comprises the following steps: S1, collecting and preprocessing corpora; S2, constructing a knowledge graph question and answer library: constructing a data set according to the data preprocessed in the step S1, and establishing the knowledge graph question and answer library by using the constructed triple data set, each triple being composed of a question entity, a question attribute and an answer; S3, constructing an entity recognition model; S4, retrieving the knowledge graph; S5, extracting keywords: performing keyword extraction on the corpus preprocessed in the step S1, and storing an extraction result into a database; S6, clusting k-means problems; and S7, similarity calculation of candidate answers: performing text similarity calculation on the candidate answers obtained in the steps S4 and S6 and the question input by the user to obtain a text answer with the highest similarity value, and outputting the text answer to the user. According to the invention, the wire incoming intention of the user can be accurately identified and corresponding knowledge can be retrieved.

The invention discloses a knowledge mapping technology-based garment customer service intelligent service method and system. A knowledge mapping database is built, then first-grade classification is carried out according to user attributes, a second-grade classification is carried out according to garment types and a third-grade classification is carried out according to garment attributes; corresponding data information is arranged under each third-grade classification; semantic parsing is conducted to visitor questions; key word information corresponding to the user attributes, garment types and garment attributes are extracted from the visitor questions; the first-grade classification, the second-grade classification and the third-grade classification corresponding to the knowledge mapping database are matched according to the extracted key word information to achieve data information corresponding to the three grades of classification; and providing corresponding answers to the visitor questions according to the achieved data information. By the use of the knowledge mapping technology-based garment customer service intelligent service method and system, data information required by a visitor can be quickly and accurately achieved, so natural language understanding and parsing capacity during intelligent customer service can be enhanced, question answering accuracy can be increased and visitor satisfaction degree can be improved.

The invention belongs to the field of video questions and answers, and particularly relates to a method for performing multi-mode video question answering by using frame-subtitle self-supervision. The method includes the following steps: extracting video frame features, question and answer features, subtitle features and subtitle suggestion features; obtaining frame features with attention and caption features with attention, and obtaining fusion features; calculating and obtaining a time attention score based on the fusion feature; calculating and obtaining the time boundary of the question by using the time attention score; calculating and obtaining answers to the questions by adopting the fusion features and the time attention scores; training a neural network by using the time boundary of the question and the answer of the question; and optimizing network parameters of the neural network, performing video question answering by using the optimal neural network, and delimiting a time boundary. The time boundary related to the problem is generated according to the self-designed time attention score instead of using time annotation with high annotation cost. In addition, more accurate answers are obtained by mining the relation between the subtitles and the corresponding video content.

The invention discloses a medical cosmetology knowledge graph automatic construction method and system and a storage medium. The method comprises a medical cosmetology knowledge graph ontology construction step, an entity extraction step and an attribute extraction step. Entity extraction is named entity extraction which is automatically identified from medical cosmetology original dialogue corpus. Attribute extraction is entity extraction, and an attribute list is constructed for entities. Attribute extraction comprises two processes of object attribute value extraction and data attribute value extraction. From original dialogue data, a shaping industry knowledge graph is quickly constructed. The invention has advantages of wide coverage and personal requirements. Through body construction, entity sampling, attribute sampling and entity alignment methods, a plastic industry knowledge graph is constructed from medical cosmetology real customer service dialogue data, and knowledge whichdoes not conform to ontology structure is manually corrected. Meanwhile, knowledge containing general errors such as sick sentences and wrongly written characters is cleaned.

The invention provides a text information emotion analysis method, a text information emotion analysis device, a computer readable storage medium and electronic equipment, and relates to the technicalfield of computers. The text information sentiment analysis method comprises the steps of determining a word vector sequence of text information according to a corpus; performing text representationprocessing on the word vector sequence in the positive sequence and the word vector sequence in the reverse sequence to obtain a plurality of first feature vectors; determining a plurality of target feature vectors corresponding to the plurality of first feature vectors according to a nonlinear mapping relationship; determining word emotion polarity corresponding to each target feature vector according to comparison between the plurality of target feature vectors and a preset emotion threshold; and determining the sentiment polarity of the text information according to the word sentiment polarity. According to the text information sentiment analysis method, the problem that the efficiency of determining the sentiment polarity is low can be solved to a certain extent, and then the efficiency of determining the sentiment polarity according to the session content is improved.

The invention provides a legal question and answer method based on a weak supervised learning and joint learning mechanism, and equipment. The method comprises the steps of obtaining legal questions input by a user; classifying the legal questions based on a joint learning mechanism by using a pre-constructed multi-label classification model and a preset legal label set to obtain legal labels of the legal questions; retrieving candidate legal question and answer pairs associated with the legal labels from a legal question and answer pair database; and outputting an answer in the at least one candidate legal question and answer pair as a reply to the legal questions. According to the legal question and answer method based on the weak supervised learning and joint learning mechanism and the equipment provided by the invention, a set of intelligent question and answer method for legal consultation is provided for an oral question and answer mode, and the defect that oral legal questions cannot be well identified and understood and intentions cannot be accurately classified by a traditional method is improved.

The embodiment of the invention discloses a customer service information processing method, a server and a system, and the method comprises the steps: receiving a customer question sent by a client, and identifying the customer demand information included in the customer question; identifying and obtaining the characteristic solving vocabularies from the customer's requirement information, determining the target requirement tags corresponding to the characteristic solving vocabularies, and determining the target requirement modification relationship among the characteristic solving vocabularies; determining target sample solving information matched with the customer's solving information from the sample solving information provided with the target solving label according to the characteristic vocabularies and the target solving modification relation; and feeding back an answer corresponding to the question sample including the target sample solving information to the client. Accordingto the technical scheme provided by the embodiment of the invention, the accuracy of answers provided for customer questions can be effectively improved.

The invention discloses an intelligent question and answer reasoning method and system based on a natural language entity relationship, and belongs to the field of natural language processing. The method the following steps: performing word segmentation and entity word extraction on each statement in a corpus; taking natural statements as edges of entity association to form entity relationships, and summarizing entity connection relationships in the corpus to form a semantic network database based on the natural language entity relationships; designing an intelligent reasoning deep learning model based on a BERT pre-training language model and a graph neural network; inputting an entity connection diagram related to questions submitted by a user into a network for reasoning, and screening the results through a multi-layer perceptron to give a final answer. Herein, the entity relation database is automatically constructed through any given natural language text corpus, entity extraction and labeling through a manual intervention means are avoided, and answers are automatically found and inferred by analyzing complex questions of the user, so that the user is helped to obtain required results more quickly and more accurately.

The invention discloses a keyword-based question and answer method and device and a medium, and the method comprises the steps: obtaining an online corpus, extracting keywords in the online corpus, screening out a target keyword matched with a training corpus from the obtained keywords, inputting the target keyword and the training corpus into a corpus training model together, and adjusting parameters of the model based on an output result of the model to obtain a semantic understanding model; after the semantic understanding model is obtained, obtaining a standard question corresponding to auser question according to the user question, thereby outputting an answer corresponding to the standard question, and completing a question and answer mode. Therefore, because the target keyword is used as an auxiliary and is used as an input sample together with the training corpus, the weight of the information matched with the keywords in the training corpus is improved, the attention of the semantic comprehension model to the keywords is actively improved, and the similar user questions are accurately recognized, so that more accurate answers can be output, and the user experience is improved.

The invention provides a dialogue reply method and device, and the method comprises the steps: obtaining the dialogue information of a user in response to a dialogue reply instruction; preprocessing the dialogue information to obtain a keyword vector corresponding to the dialogue information; inputting the keyword vector into a pre-trained response strategy classification model to obtain a response strategy classification result output by the response strategy classification model; combining the keyword vector with the response strategy classification result to obtain a target input vector; inputting the target input vector into a pre-trained reply model to obtain reply information corresponding to the dialogue information; and feeding back the reply information to the user. By applying the dialogue reply method provided by the invention, the dialogue information of the user can be accurately replied, so that the information required by the user can be provided for the user.

The invention provides a machine reading understanding model based on knowledge graph gains, which is used for receiving a text data set including text documents and questions and a vocabulary automatically generated according to the text data set and obtaining answers to the questions according to the content of the text documents. The machine reading understanding model comprises a document question arrangement module, a named entity recognition module used for performing named entity recognition processing on the text data set; an ERNIE context language module; an external knowledge base which comprises a WordNet knowledge base and a ConceptNet knowledge base and is used for receiving the vocabulary and correspondingly generating a WordNet knowledge feature vector and a ConceptNet knowledge feature vector; a knowledge matching and connecting layer which is used for connecting corresponding word vectors with WordNet knowledge feature vectors and ConceptNet knowledge feature vectors for entities which are successfully matched in text documents and questions; an attention calculation unit which is used for performing bidirectional attention operation and self-attention operation on the vectors correspondingly to obtain answers; and a result generation unit which is used for receiving and judging the output answer.