253 results about "Question Text" patented technology

The invention relates to a deep learning-based question and answer matching method. The method comprises the following steps of: 1) sufficiently learning word orders and sentence local features of a question text and an answer text by utilizing two underlying deep neural networks: a long short-term memory network LSTM and a convolutional neural network CNN; and 2) selecting a keyword with best semantic matching on the basis of a pooling manner of an attention mechanism AM. Compared with existing methods, the method has the advantages of being in low in feature engineering workload, strong in cross-field performance and relatively high in correctness, and can be effectively applied to the fields of commercial intelligent customer service robots, automatic driving, internet medical treatment, online forum and community question answering.

The present application relates to an intelligent question answering method, apparatus, computer device and storage medium. The method comprises the following steps of: acquiring question text of a target domain and extracting a target entity from the question text; obtaining The word vector of the target entity by using the pre-constructed word vector model. inputting The word vector into the pre-trained user intention classification model, and acquiring the user intention types of the question text. Generating A structured query statement according to the user intention type and the target entity, and finding the matching question answers from the knowledge base of the target domain according to the structured query statement. The method realizes intelligent question answering based on natural language processing technology and database technology, which can effectively improve the accuracy of question text matching with the answers in the knowledge base, and improve the accuracy ofintelligent question answering.

The invention discloses an automatic question-answer processing method and an automatic question-answer system. The method includes: acquiring question text from question-answer data pairs collected in advance, performing word separation on the question text to obtain the corresponding key words of the question text, and building the index relation between the key words and the question text; whenoptional target question text is received, and performing word separation on the target question text to acquire target key words corresponding to the target text question text; according to the index relation of the key words and the question text, determining key words matched with the target key words, and acquiring the question text having index relation with the key words to serve as the candidate question text; calculating the semantic similarity of the candidate question text and the target question text; determining an answer corresponding to the target question text according to thesemantic similarity. The method has the advantages that the semantic similarity of the target question text and each question text is considered to determine the answer of the target question text, and the accuracy of automatic question-answer processing is increased.

The invention relates to the technical field of man-machine interaction, and discloses an information processing method and device for realizing intelligent question answering. The information processing method comprises the following steps of: carrying out sentence segmentation on question text information to obtain a user question; and searching a standard question most similar to the user question and corresponding answer information from a QA library on the basis of a question similarity. Compared with the existing keyword retrieval-based question answering method, the method disclosed by the invention does not need to require the users to have keyword decomposition ability, is automatic in the whole process and is capable of greatly enhancing the user experience and improving the search effect and the pertinence and effectiveness of answers. Meanwhile, through fusing natural language understanding technologies such as sentence model analysis, lexical analysis and lexical meaning extension, and carrying out comprehensive calculation on multi-dimensional similarity, the method is capable of improving the correctness of a final sentence similarity in a Chinese automatic question answering process, and enabling a Chinese intelligent question answering system to be possible.

The invention discloses a deep learning-based question classification model training method and apparatus, and a question classification method and apparatus. The question classification model training method comprises the steps of extracting feature information samples in question text samples, and generating corresponding first eigenvector samples; performing spatial transformation on the first eigenvector samples to obtain second eigenvector samples; inputting the second eigenvector samples to a plurality of convolutional layers and a plurality of pooling layers in a multilayer convolutional neural network, and by superposing convolution operation and pooling operation, obtaining first fusion eigenvector samples; inputting the first fusion eigenvector samples to a full connection layer in the multilayer convolutional neural network to obtain global eigenvector samples; and training a Softmax classifier according to the global eigenvector samples to obtain a question classification model. The method can avoid a large amount of overheads of manual design of features; and through the question classification model, a more accurate classification result can be obtained, so that locating of standard question and answer is improved.

The invention provides a method and a device for matching between questions and answers and relates to the technical field of intelligent questions and answers. The method for matching between questions and answers include extracting keywords input into a question text, determining a target matching question text from a pre-built question base by an index filtering method according to the keywords, determining the optimum matching question text with highest similarity with the input question text from the target matching question text on the basis of the Levenshtein distance algorithm, and outputting an answer text corresponding to the input question text according to the optimum matching question text. By the method and the device, answers matched with questions can be outputted and inputted within a short time, thus, time for matching between the question and the answer can be shortened while matching accuracy can be improved.

An embodiment of the invention discloses a problem text generation method and device, equipment and a medium. The method comprises the following steps: the embodiment of the invention determines the coded hidden layer state vector of each input word by adopting an encoder based on the attribute parameters of each input word; for each candidate word, the decoder and an attention mechanism model areused to determine the context vector and decoded hidden layer state vector of each candidate word as the current output word; and for at least one generation mode, the probability value of each candidate word as the current output word and the weight value of the generation mode are respectively calculated according to the context vector and the decoded hidden layer state vector, and the final probability is further determined to screen the current output word among the candidate words based on the final probability. The technical scheme improves the accuracy and diversity of generating the question text based on the answer text.

The invention belongs to the technical field of internet databases, and particularly relates to a community question and answer system and method based on multi-task learning and an electronic device.The system comprises an answer selection model training module which is used for inputting answer input and question input into a bidirectional long-short memory network for encoding, inputting the encoded answer input and question input into a multi-dimensional attention layer, flattening and connecting an output result, and calculating the loss of a prediction result and a real result; a question classification model training module which is used for inputting questions into a bidirectional long and short memory network for encoding, inputting the encoded questions into a two-layer full connection network, and calculating the loss of a prediction result and a real result through a softmax layer; and a joint training module which is used for unifying the answer selection task and the question text classification task under a loss function for joint training to obtain an answer related to the input question. According to the application, the accuracy of the forum community question and answer system can be improved, and the search efficiency of the user is improved.

Disclosed is an automatic labeling method for mathematic question knowledge points. The labeling method is based on a depth learning model. The labeling method comprises the steps of S1, preprocessinga mathematic question text, wherein standardization, word segmentation and stop word removal processing are performed separately; S2, preparing a mathematic field corpus, and training a mathematic field word vector according to the character level and the word level; S3, setting up a two-layer Text CNN neural network model, inputting a mathematic question training sampling set with labels, and training the neural network model. The knowledge points of mathematic question in a novel question bank can be automatically labeled by means of the model.

The invention relates to a shallow layer model and depth model combination-based question text classification method, and belongs to the technical field of computer natural language processing. The method comprises the following steps of: firstly extracting a feature word set of a question text, obtaining corresponding feature word weights by utilizing normalized word vectors after vectorization,and taking the corresponding feature word weights as a part of input of a shallow layer linear model; convoluting the question text by a convolutional network by using multiple convolution cores withdifferent window sizes, rearranging the feature vectors extracted by different convolution cores with same-length convolution windows, respectively inputting the feature vectors into corresponding recurrent neural networks, and finally taking syntax semantic feature vectors, obtained by linking outputs of the plurality of recurrent neural networks together, of the question as another part of inputof the shallow layer linear model; and finally obtaining a classification result of the question by the shallow layer model according to an input spliced by the feature word vectors and an output ofa depth model. The method is capable of overcoming the defects of single depth models and effectively enhancing the question classification correctness.

The invention relates to a keyword extraction method and a keyword extraction system. The keyword extraction method and the keyword extraction system of the invention are used to solve the technical problem that the keywords expressed by correct semantics cannot be accurately obtained. The keyword method comprises the following steps: vectorizing a question text to form a question corpus including the vector features of keywords; and extracting the keywords in the question corpus by utilizing BLSTM RNN (Bidirectional Long Short Term Memory Recurrent Neural Network).

The invention discloses a Bi-LSTM label recommendation method based on an attention mechanism. The Bi-LSTM label recommendation method based on the attention mechanism comprises the following steps: collecting an experimental data set; analyzing specific text data from the experimental data set; preprocessing the text data; extracting semantic features from the preprocessed question text description; constructing a multi-label classification model; recommending a proper label for the new problem through the constructed multi-label classification model; and evaluating and analyzing a label recommendation result. The method has the beneficial effects that the label recommendation task is mainly converted into the multi-label classification problem through the Bi-LSTM model based on the attention mechanism, the labels are automatically recommended according to the text description content of the problem, and the label recommendation accuracy is improved.

The invention discloses an automatic elementary mathematic algebra type question answering method and system. The method comprises the following steps of: question input; question understanding: classifying each category of elementary mathematic algebra type questions and carrying out word segmentation and part-of-speech labelling on question texts; extracting a direct statement mathematic relationship by using a syntax-semantic hybrid model and adding an implicit type mathematic relationship according to question types so as to form an algebra relationship group; machine solution: distributing variables for entities in the formed algebra relationship group, converting the algebra relationship group into an algebra equation group, obtaining an entity-variable comparison table, and automatically solving the algebra equation group by a machine; and quasi-man answer generation: recovering semantic meanings of the variables in the algebra relationship group solution process according to solution sequences of the variables and the entity-variable comparison table, and combining the question texts to form a quasi-man answer process. According to the method and system, the automation degree of answering mathematic algebra type questions can be greatly improved.

The invention provides a problem generation method and device, equipment and a storage medium. The method comprises: encoding a first word vector corresponding to a reference text, an answer information vector and a text position vector corresponding to the current question and answer round number through a first coding model, and obtaining a first semantic vector sequence; encoding a second wordvector corresponding to the historical question and answer text through a second encoding model to obtain a second semantic vector sequence; decoding the first semantic vector sequence and the secondsemantic vector sequence through a decoding model to obtain a problem text corresponding to the current round of number; outputting the question text. Through the method and the device, the problems of coherence and good dialogue performance can be generated by combining historical dialogue contents.

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.

A conversational question generation system dynamically generates conversational questions for insurance claim processing. The conversational question generation system includes various modules and graphical user interfaces that provide a streamlined mechanism for creating new conversational questions for insurance claim processing. The conversational question generation system may include various levels of usability that distinguish between a question programmer of the conversational question generation system and an insurance claim agent that uses the created conversational questions. In generating conversational questions, the conversational question generation system may include graphical user interfaces directed to the question details, the answer details, the question text, or other aspects of the conversational questions.

The present disclosure provides a semantic matching method and device of a question and answer text, which includes the steps of obtaining a feature vector sequence with contextual local features of aquestion text and a feature vector sequence with contextual local features of a candidate answer text by using a loop neural network; based on an attention weight of each feature vector in a featurevector sequence having contextual local features of the question text and the candidate answer text and a feature vector sequence having contextual local features of the question text and the candidate answer, generating a feature vector sequence with contextual local features and global features of the problem text and a feature vector sequence with contextual local features and global features of the candidate answer text; and according to the feature vector sequence with local and global contextual features of the question text and the feature vector sequence with local and global contextual features of the candidate answer text, determining the semantic matching degree between the question text and the candidate answer.

The present disclosure discloses an automatic question-answer (QA) processing method and an automatic QA system. The method includes: obtaining, after receiving a target question text, a target keyword corresponding to the target question text; determining a candidate question text that matches the target keyword; calculating a semantic similarity value between each candidate question text and the target question text; and determining, based on the semantic similarity value, an answer corresponding to the target question text. In the present disclosure, a semantic similarity between a target question text and each question text is considered, to determine an answer to the target question text, improving accuracy of automatic QA processing.

A medical decision system includes a patient information database including patient information, a questions database including a plurality of questions to be presented to a user, each having question text associated therewith and a predetermined question code associated with the question text, an answers database including a plurality of answers each having answer text associated therewith and a predetermined answer code associated with the answer text, and a plurality of medical related information databases including medical related information. At least one of the question text and answer text are locked to prevent alteration thereof.

The invention discloses a consultation service-based information processing method and apparatus. The method comprises the steps of receiving a question input by a user; performing text format processing on the question to generate a standard question text; calling various started consultation engines to process the standard question text to generate initial consultation results; according to a preset reliability rule, determining the reliability of the initial consultation results; according to the determined reliability of the initial consultation results, determining the initial consultation result matched with the standard question text; and displaying the initial consultation result as a consultation result of the question. Through the method, for diversified user questions, answers better meeting user expectation can be selected out from multiple consultation results, so that the accuracy of reply is improved.

The invention relates to the field of natural language processing, and a model based on a self-attention mechanism is used to extract attention features between questions and answers so as to obtain the matching degree between the questions and the answers according to the attention features. The invention specifically discloses a question and answer matching processing method and device, a question and answer matching model training method and device, equipment and a storage medium. The method comprises the steps of obtaining a question text and an answer text; performing word segmentation processing on the question text and the answer text to obtain corpus word segmentation data; performing embedding processing on the corpus word segmentation data to obtain embedded representation data;based on a feature extraction sub-model, performing feature extraction on the embedded representation data to obtain a self-attention feature vector, and the feature extraction sub-model being the model based on the self-attention mechanism; and based on a matching sub-model, generating question and answer matching data according to the self-attention feature vector, and outputting the question and answer matching data.

The invention discloses an answer search method and device based on semantic indexing, electronic equipment and a storage medium. The method comprises the steps of obtaining a question text input by auser; carrying out vector conversion on the problem text according to a pre-trained semantic index model to obtain semantic vector expression of the problem; matching the semantic vector expression of the question with each answer vector expression in a pre-established answer vector index library; wherein the answer vector index library is constructed by converting all answers in an answer samplepool into vector expressions according to a semantic index model; and obtaining a corresponding answer text according to the matched answer vector expression, and providing the corresponding answer text as a search result for the user. The method can solve the technical problem that in the prior art, a common similarity matching technology only simply matches questions from a character surface, so that only answers related to some character surfaces can be obtained, the answer screening time of a user is saved, and the use experience of the user is improved.

The invention discloses an answer acquisition method and device. The method comprises the following steps: acquiring a question text; determining whether the question type of the question text is matched with the answer type in a preset answer type library or not through a pre-trained question classification model; and if the question type of the question text is not matched with the answer type in the preset answer type library, obtaining the answer of the question text according to at least one semantic similarity between the first deep semantic vector of each answer in the preset answer library and the second deep semantic vector of the question text. The method is applied to finance technology (Fintech). A pre-trained question classification model is used; according to the method, whether the question type of the question text is matched with the answer type in the preset answer type library or not is determined, and when the question type of the question text is not matched with the answer type in the preset answer type library, the semantic matching degree of the question text and each answer in the preset answer library is represented by at least one semantic similarity, so that the answer obtaining accuracy is improved.

The present invention discloses a method for constructing a deep visual Q&A system for visually impaired persons. In the training phase, the method comprises: taking collected pictures and a corresponding Q&A text to constitute a training set; extracting picture features for the pictures by using the convolutional neural network; for a question text, converting questions into a word vector list by using the word vector technique, and taking the word vector list as input of the LSTM so as to extract question features; and finally, after carrying out element dot product on the pictures and the question features, carrying out classification on the pictures and the question features so as to obtain an answer prediction value, comparing the answer prediction value with an answer tag in the training set, calculating the loss, and using the back propagation algorithm to optimize the model. In the running phase, the method comprises that: a client obtains photos taken by the user and the question text, and uploads the photos and the question text to a server; the server inputs the uploaded photos and question text into a trained model, extracts question features by using the same manner, outputs a corresponding answer prediction value by using a classifier, and returns the answer prediction value to the client; and the client returns the answer prediction value to the user in a form of voice input.

The invention discloses an image question and answer method based on multi-objective association deep reasoning. The method comprises the following steps of 1, carrying out data preprocessing on an image and a text described by a natural language of the image; 2, carrying out attention mechanism reordering on each target based on an adaptive attention module model enhanced by geometric features of a candidate box; 3, constructing a neural network structure based on an AAM model; and 4, model training: training neural network parameters by using a back propagation algorithm. The invention provides a deep neural network for image question answering, in particular to a method for performing unified modeling on image-question text data, performing reasoning on each target feature in an image, and reordering attention mechanisms of the targets so as to answer questions more accurately, and a better effect is obtained in the field of image question answeringThe invention discloses an image question and answer method based on multi-target association deep reasoning. The method comprises the following steps: 1, carrying out data preprocessing on an image and a text described in a natural language of the image, and 2, carrying out attention mechanism reordering on each target based on an adaptive attention module model with enhanced geometrical characteristics of a candidate box. And 3, a neural network structure based on an AAM model. And 4, model training: training neural network parameters by using a back propagation algorithm. The invention provides a deep neural network for image question answering, and particularly provides an image-image question answering method. According to the method, the data of question texts are subjected to unified modeling, reasoning is carried out on the characteristics of all the targets in the image, attention mechanisms of all the targets are reordered, so that questions are answered more accurately, and a good effect is obtained in the field of image questions and answers.

The invention discloses a subjective and objective classifier building method and system. The subjective and objective classifier building method and system are characterized in that emphasis is put on training questions and answers, base classifiers are built in terms of the question text and the answer text and then are infused, and a final subjective and objective classifier is obtained. Therefore, the answer classification is added in the subjective and objective classification, the question classification is corrected and calibrated by combining the answer features, and therefore subjective and objective classification based on question and answer complementation is achieved, the shortcoming of low accuracy of the classifier caused by ambiguity of a question training sample is overcome, the accuracy of classifying questions by aid of the subjective and objective classifier is improved, and further the performance of a question and answer system is improved.

The invention relates to the field of intelligent decision making and discloses a consultation request processing method and device, a computer equipment and a storage medium, to reduce question and answer interaction between an operator and a user, reduce the situation of low efficiency caused by human participation and effectively improve consultation feedback efficiency. The method comprises the following steps: obtaining a consultation request triggered by a user through an application, wherein the consultation request comprisingto-be-answered consultation text information; determining whether the to-be-answered consultation text information being questioning text information or not through a preset text classification model; if being determined that the to-be-answered text informationbeing question text information, to extract a to-be-answered question from the to-be-answered text information; determining the similarity between the to-be-answered question and the historical questioning question according to a preset similarity model; determining target answering content according to the similarity between the to-be-answered question and the historical questioning question; and performing consultation feedback on an answer corresponding to the consultation request.

The invention discloses a question extraction method and electronic equipment. Key features in each dialogue corpus are determined by aiming at the conversations, which aim at different questions, ofusers with a customer service staff; according to the key features, a real question text is determined; according to different dialogues, semantic features and structural features in the dialogue aredetermined; and therefore, the problem that the authenticity of the user can not be accurately and quickly determined when a keyword list is not complete since keywords in the keyword list are extracted in the dialogue can be solved.

The invention discloses a call center service management system, and relates to the technical field of telephone communication services. The system comprises a database, a recording module, a voice conversion module, a keyword extraction module, a matching module, a matching degree identification module and a sorting module. The recording module records audio signals of each call between customerservice and user, converts the recorded audio signals into user question text and customer service answer text, and stores the recorded audio signals in the database; By selecting the preset problem which is closest to the user problem through the matching module, the matching degree recognition module compares the customer service answer text with the keywords of the standard answer to obtain thematching degree result, The ranking module ranks the customer service answers to the same preset questions according to the matching results, and the preset questions and the corresponding customer service answers to the preset questions are stored in the database, thus forming a case base for customer service novices to learn.