1100 results about "Intent recognition" patented technology

The apparatus for foreign language study includes: a voice recognition device configured to recognize a speech entered by a user and convert the speech into a speech text; a speech intent recognition device configured to extract a user speech intent for the speech text using skill level information of the user and dialogue context information; and a feedback processing device configured to extract a different expression depending on the user speech intent and a speech situation of the user. According to the present invention, the intent of a learner's speech may be determined even though the learner's skill is low, and customized expressions for various situations may be provided to the learner.

The invention relates to a robot man-machine interaction method based on user mood and intension recognition. User mood recognition is carried out by combining human body biological information such as breath, the heart rate and the dermal electricity with facial expression recognition. The user intension is recognized through different kinds of physical sensor information such as pressure, photoelectricity and temperature. Intelligent control and decision making are carried out according to the user mood and intention recognition results. A corresponding execution mechanism of a robot is controlled to complete limb movements and voice interaction. By means of the robot man-machine interaction method based on user mood and intension recognition, the robot can understand the user intension more thoroughly, know the metal changes of a user, meet the functional requirement for emotion care for the user, and better participate in the life of users such as the old people and children.

A method and intent recognition system that triggers voice recognition is provided. The intent recognition system includes an intent recognition device that detects a user location and / or behavior in a building and detects a triggering event based on the user location and / or behavior that suggests the user wants to input a service request, a prompt device that is triggered to provide a prompt based on the detecting by the intent recognition device and the detected triggering event, a listening device that receives an auditory input from the user, a signal processing device that processes the received auditory input and generates a control signal, and in-building equipment that receives the control signal from the signal processing device and controls the in-building equipment based on the control signal.

The invention provides an intelligent interaction method, an electronic device and a storage medium. The method comprises the following steps of: previously constructing dialogue process templates ofdifferent business processes and training an intention recognition and classification model; after a statement input by a user is received, executing a series of processing such as preprocessing, statement type recognition and emotion analysis on the statement, recognizing an emotion category of the user, and recognizing a user intention from the processed statement by utilizing the pre-constructed dialogue process templates and the trained intention recognition and classification model; and querying a corresponding knowledge base according to the recognition user intention, combining the query result with the emotion category of the user to generate a reply, and feeding back the reply to the user. By utilizing the method, the correctness of intelligent question answering systems can be improved, users can be guided to execute process dialogues according to input of users, and users can be initiatively guided to complete online business processes.

Embodiments of the invention provide a voice search method and device and electronic equipment, and relates to the technical field of audio processing. The method comprises the following steps of: receiving a to-be-identified voice; carrying out intention identification on the to-be-identified voice so as to obtain a search intention of a target user that issues the to-be-identified voice; obtaining a vocal print feature of the to-be-identified voice and taking the vocal print feature as a to-be-identified vocal print feature; identifying the target user through the to-be-identified vocal print feature; and carrying out search on the basis of the target user by utilizing the search intention so as to obtain a search result. By using the method provided by the embodiments to carry out voice search, the correctness of voice search results can be improved.

The invention discloses a method and system for intention recognition based on a context. The method comprises the steps of carrying out manual annotation on problems provided by a user and included in a user historical consulting log on the basis of the user historical consulting log; extracting features of each problem to generate a training corpus; carrying out model training on the generated training corpuses according to a supervised learning algorithm to obtain a training model; estimating the current problems and obtaining an estimated user intention recognition result on the basis of the obtained training model.

The invention discloses an intelligent customer service response method and system, relates to the technical field of intelligent customer service, and can solve the problems of low response efficiency and poor user experience caused by relying on manual customer service in the prior art. The method comprises the steps of preprocessing the customer service historical chat statements to export customer service corpora, and conducting the word segmentation training on the customer service corpora to obtain a language recognition model; carrying out intention identification on corpora asked by auser to match a corresponding service module, carrying out retrieval analysis in an elasticsearch database corresponding to the service module by utilizing an intention identification result, and summarizing output associated keyword answers to form a pre-selected data set; taking a text formed after word segmentation of the question corpus of the user as an input, and carrying out Elasticsearch retrieval in a pre-selected data set to output candidate keyword answers; converting the segmented keywords and the candidate keyword answers into word vectors respectively, calculating the similaritybetween the segmented keywords and the candidate keyword answers in pairs by using a WMD algorithm, and selecting the keyword answers based on the similarity values to respond to the user.

The invention discloses an artificial intelligence-based spoken language query identification method and apparatus. The method comprises the steps of training a convolutional neural network according to a query field annotated by a spoken language retrieval corpus to generate a retrieval field identification model; and training a recurrent neural network according to a query intention corresponding to the query field and annotated by the spoken language retrieval corpus, and parameter information to generate a retrieval intention identification model corresponding to the query field. According to the method and the apparatus provided by embodiments of the invention, the high-applicability and high-automation retrieval field identification model and retrieval intention identification model are generated through training, and the spoken language query intention of a user and the parameter information corresponding to the spoken language query intention can be accurately obtained, so that the efficiency and accuracy of spoken language query identification are improved.

The invention relates to an electroencephalographic and electromyographic information automatic intention recognition and upper limb intelligent control method and system, which are used for rehabilitation treatment of the upper limb of a stroke patient, an electroencephalographic and surface electromyographic signal collector collects and processes the electroencephalographic and surface electromyographic signals of the patient in real time, a mixed kernel function formed by weighting a polynomial kernel function and an RBF kernel function weights is used to perform fitting and prediction, soas to more accurately identify and monitor the motion intention of the patient, and judge the corresponding degree of rehabilitation, according to which a corresponding rehabilitation training strategy is adopted. When the rehabilitation degree of the upper limb of the stroke patient is low, passive training control is adopted. When the rehabilitation degree of the upper limb of the stroke patient is high, active, assisted and resistive control modes are adopted. The hybrid kernel function support vector machine model provided by the invention has better learning ability and generalization performance, high prediction accuracy and good control performance, and the prediction result meets the index requirements of a rehabilitation robot for stroke patients.

The invention relates to a brain electric and eye electric signal decision fusion method for identifying control intention, comprising the following steps: extracting relative wavelet packet energy of pre-processed EEG (electroencephalogram) as motor imagery feature and taking typical variable of pre-processed EOG (electrooculogram)as motor imagery feature; applying the constructed state classifier based on the motor imagery feature of the EEG to obtain state classification result; and applying the constructed motor imagery multi-class classifier based on the motor imagery feature of the EOG to obtain motor imagery classification result; finally, and using the state classification result obtained by the EEG and the motor imagery classification result obtained by the EOG to perform decision fusion so as to obtain identification result of the control intention represented by the motor imagery. The method of the invention can effectively solve interference for identification result of the control intention caused by eye movement irrelevant to the control intention and improve accuracy of the motor imagery identification and reliability of the control intention information.

The invention relates to the technical field of computers, and provides a realization method and apparatus for an intelligent question and answer system. The method comprises the steps of obtaining question contents of a first user, and extracting key data in the question contents; according to the key data, generating template questions, and dividing the template questions in a preset question intention; obtaining the question contents currently input by the user, and according to the key data carried in the current question contents, determining that slot position information is missing in the current question contents; determining the question intention of the current question contents; from the template questions comprised in the corresponding question intention, confirming at least one template question meeting a condition that a matching degree between the at least one template question and the current question contents reaches a preset threshold; and filling the slot position information with key information in the at least one template question to obtain current question contents filled with a version. The input question intention identification and conversation state tracking functions are realized, so that the system can generate corresponding answers in combination with contexts.

The present invention discloses a state machine based context-sensitive multi-round dialog management system, comprising: an input module, for receiving multi-modal input information from a user; an intention identification engine module, for identifying intention information in the multi-modal input information; an intention module, for bringing multiple intention information identified by the intention identification engine module into one-to-one correspondence with multiple intention sub-modules at back ends; a state machine module, comprising a plurality of state machines for managing a relevant context in the dialog management system and providing support for an output result; an instruction parsing engine module, comprising a plurality of instruction parsing engine sub-modules for parsing corresponding intention information and acquiring the parsed multiple intention information; and an output module, for acquiring policy information according to the results from the parsing engine module and the intention identification module, and transmitting the policy information to the state machine module.

The embodiment of the invention provides a method and a device for identifying query intention based on a mixed strategy. The method comprises the following steps: acquiring a query sentence of a user; obtaining a first candidate query intention by using a semantic rule template according to a query statement of a user; according to the user's query statement, using the machine learning model to obtain the second candidate query intention; matching the first candidate query intent with the second candidate query intent; if the first candidate query intent and the second candidate query intentmatch successfully, outputting as semantic query intent. The first candidate query intent, the second candidate query intent, and the semantic query intent are at least one. The embodiment of the invention combines the semantic rule template and the machine learning model to carry out query intention identification, can better understand the query intention of a user's query sentence, and improvesthe semantic identification ability and the understanding effect.

The invention provides an intention recognition method and an intention recognition device under a dialogue framework. The method comprises the steps: obtaining the matching degree of a to-be-recognized corpus and rule monomers in a preset rule template, wherein the preset rule template comprises a plurality of rule monomers, and each rule monomer corresponds to one label; judging whether a matching degree greater than a preset threshold exists or not; if so, taking the label of the rule monomer corresponding to the maximum matching degree as an intention recognition result; if not, inputtingthe corpus to be recognized into a pre-trained intention recognition model. Moreover, the output of the pre-trained intention recognition model is used as an intention recognition result. When the rule template cannot effectively recognize the user intention, the machine learning model is used for recognizing the user intention. The user intention can be accurately recognized. The requirement forthe number of training samples is low. The development of artificial intelligence equipment such as a question-answering system and a dialogue robot is promoted.

The invention provides an entity labeling method, an intention identification method and a corresponding device, and a computer storage medium. The entity labeling method comprises the following steps: encoding the attribute labels of at least part of the words in a sentence by using a knowledge map to obtain a first expression vector of at least part of the words; Encoding at least part of the words in the sentence based on the sentence structure to obtain a second expression vector of at least part of the words; The first expression vector and the second expression vector are fused to obtainan entity labeling result of the sentence. The intent recognition method comprises the following steps: the attribute tags of at least part of the words in a sentence are combined and encoded by using a knowledge map to obtain a first sentence vector of the sentence; Encoding the sentence based on the sentence structure to obtain a second sentence vector of the sentence; The first sentence vectorand the second sentence vector of the sentence are fused to obtain an intention recognition result of the sentence. The method provided by the invention can improve the accuracy of entity labeling and intention identification.

The invention provides a method for identifying user query intention. The method comprises the following steps that user query to be identified is divided into a named entity e1, a named entity e2 and a named entity context ct; probability p (r Iotae 1, e2, ct) of a latent topic r is calculated in the user query under the condition that the named entity e1, the named entity e2 and the named entity context ct are included according to a set and a topic model of a user query log, a topic r which makes the probability p (r Iota e1, e2, ct) be maximum is found out, and the topic r is used as the latent relation between the named entity e1 and the named entity e2. The topic model is obtained by query log aggregation document set training, and a query log aggregation document set is a document set about combination of the two-two named entities formed by aggregating the context of every two named entities in the query log. The method for identifying the user query intention is high in accuracy rate and recalling rate of identification, low in time complexity, low in load to a search engine, capable of improving user experience, and especially suitable for searching of a mobile terminal or for local application searching of the mobile terminal.

The invention discloses an NAR (Nonlinear Autoregressive Models) neural network vehicle speed prediction method based on driving intention recognition. The method comprises the following steps of driving intention classification and recognition parameter selection; fuzzy reasoning recognition of the driving intention; NAR neural network off-line training; and NAR neural network on-line vehicle speed prediction: firstly performing driving intention recognition, and then inputting the driving intention obtained through recognition and the vehicle speed time sequence into an NAR neural network together so as to realize the vehicle speed prediction of the vehicle in a period of time in future. The NAR neural network vehicle speed prediction method has the advantages that the NAR neural network is used for performing vehicle speed prediction; the neural network input includes the network output feedback; the method is suitable to be used for solving the nonlinear problem on the time sequence; and the multi-step prediction precision can be obviously improved. The driving intention time sequence and the vehicle speed are introduced to be jointly used as the input; the fuzzy reasoning is used for analyzing the pedaling operation of a driver; the expectation of the driver on the future change trends of the vehicle speed is sufficiently shown; and the vehicle speed prediction accuracy is improved.

The invention discloses a natural language understanding method based on a pre-training model. The method comprises the following steps: establishing a pre-training model based on a bidirectional depth Transformer; carrying out word segmentation processing on the to-be-understood statement, and respectively adding special tags at the beginning and the end of the to-be-understood statement to obtain a text vector of the to-be-understood statement; taking the text vector of the to-be-understood statement as input, and calling the pre-training model to obtain a text semantic vector of the to-be-understood statement; performing intention recognition; and performing entity identification. According to the method, intentions can be accurately and comprehensively understood, entities can be recognized, and a solid foundation is provided for subsequent conversations. The quality and the user experience of the man-machine conversation system can be remarkably improved.

The invention discloses an intelligent vehicle safety situation assessment method considering multi-vehicle interaction. The method comprises the following steps:1, establishing an intention recognition module through a long-short-term memory network; 2, defining logic judgment based on probability maximum classification through the intention recognition module according to the historical track information of the vehicle, and calculating the driving intention probability of the surrounding vehicles of the predicted vehicle; 3, outputting a potential risk by adopting a risk assessment module based on a driving safety field through analyzing the interaction among the driver, the vehicle and the road; 4, establishing a comprehensive situation evaluation model in combination with the driving intentions, calculated by the risk evaluation module and the intention recognition module, of the surrounding vehicles of the predicted vehicle, and outputting a dynamic potential risk map consideringmulti-vehicle interaction in a dynamic traffic scene through the comprehensive situation evaluation model; according to the invention, the dynamic potential risk map considering multi-vehicle interaction can be output, and the calculated risk map allows the intelligent vehicle to evaluate the driving state of the intelligent vehicle by using the real-time risk value so as to perform early warningand take appropriate measures.

The invention discloses an intelligent household system and a data processing method thereof. The intelligent household system comprises an input system for obtaining user control information, an intelligent analysis system for performing intention identification on the user control information so as to obtain potential intentions, an intelligent control system for determining one clear intention from the potential intentions according to intelligent equipment information fed back by intelligent equipment, an equipment control system for converting the clear intention into an operation instruction, and the intelligent equipment executing corresponding motion according to the operation instruction. Compared to a technical scheme in which users are required to remember specific instructions in the prior art, the intelligent household system has the following advantages: the user friendliness and intelligentization are higher, user experience is greatly improved, and promotion and popularization of the intelligent household system are facilitated.

The invention provides a voice query method and device, a storage medium and terminal equipment. The method comprises the steps that a voice query request of a user is received; voice recognition is performed on the voice query request, and a query statement is acquired; intention recognition and object recognition are performed on the query statement to obtain a query intention and a query objectterm of the query statement, wherein the query object term corresponds to at least one query object; the query object of the query statement is determined based on the query object term; and a retrieval database is queried according to the query intention and the query object of the query statement to obtain a query result. By the adoption of the voice query method and device, the storage mediumand the terminal equipment, query accuracy can be improved.

The embodiment of the invention discloses a personalized recommendation method and device, a server and a medium, wherein, the method comprises the following steps: according to a user portrait, usinga pre-trained demand intention model to predict a demand user group with a demand for a recommendation object from an original user group; utilizing the pre-trained intention recognition model and the user portrait of each user in the demand user group, the behavior intention of each user is predicted. The behavior intention of each user is matched with the attribute characteristics of the recommendation object, and the personalized recommendation is made for each user according to the matching result. The embodiment of the invention solves the problem of low accuracy of personalized recommendation for users, improves recommendation accuracy, further improves response rate of users after recommendation, and improves conversion rate of commodities or services.

The invention discloses an intention recognition method and device applied to an intelligent customer service robot, and belongs to the technical field of artificial intelligence. The method includesS0, acquiring the dialogue text of a user; S2, judging whether the dialogue text contains intention or not, if so, executing the step S4, if not, ending processing, and if not, executing the step S3;S3, performing context expansion on the dialogue text, and after the step S3, executing the step S4; S4, identifying a named entity set in the dialogue text, and determining intention knowledge pointsassociated with the named entity set; S5, expressing the dialogue text by adopting a distributed word vector, and predicting by using a plurality of pre-trained semantic classification models to obtain a plurality of pieces of semantic information; and S6, combining and optimizing the intention knowledge points and the semantic information by using an Ensemble framework to obtain the intention ofthe user. According to the embodiment of the invention, the intelligent customer service robot can quickly and accurately identify the intention of the user, and guarantee is provided for the robot to accurately answer the user question.

The embodiment of the invention provides an intelligent question-answering method and device based on a knowledge graph. The intelligent question-answering method comprises the steps: receiving a userquestion transmitted by a client, and respectively carrying out the entity recognition and intention recognition; and generating a knowledge graph query statement according to the entity identification result and the intention identification result, searching in a knowledge graph database to obtain an answer corresponding to the user question, and returning the answer to the client, wherein the knowledge graph database comprises entity names, entity attribute information and entity relationship information. The intelligent question-answering method and device based on a knowledge graph generates knowledge graph query statement after entity identification and intention identification are carried out based on a user question, and carries out searching in the knowledge graph database containing the entity name, the entity attribute information and the entity relation information to obtain the answer corresponding to the user question and return the answer to the client, thus achieving intelligent automatic answering of the user question, and improving the answering efficiency and accuracy.

The invention relates to a driver lane change intention identification method based on a hidden Markov model. The driver lane change intention identification method based on the hidden Markov model isbuilt through analysis on the running state of a car and the running states of surrounding vehicles, and that is to say, through the observation states of observable lane change characterization parameters, hidden driver lane change intention is predicted. On the basis of a highway vehicle database which centralizes NGSIM data, lane change characterization parameter samples are extracted, the hidden Markov model is trained and verified with an HMM toolbox editing algorithm in MATLAB, and finally a prediction result of the driver lane change intention has high accuracy.

A system and method of creating the natural language understanding component of a speech / text dialog system. The method involves a first step of defining user intent in the form of an intent flow graph. Next, (context, intent) pairs are created from each of the plurality of intent flow graphs and stored in a training database. A paraphrase task is then generated from each (context, intent) pair and also stored in the training database. A zero-shot intent recognition model is trained using the plurality of (context, intent) pairs in the training database to recognize user intents from the plurality of paraphrase tasks in the training database. Once trained, the zero-shot intent recognition model is applied to user queries to generate semantic outputs.

An intelligent surveillance system is disclosed for the identification of suspicious behavior near the exterior of a vehicle. The system of the invention is comprised of a “fish-eye” visible camera imaging system installed on the interior ceiling of an automobile for the 360-degree imaging and observation of the lower hemisphere around the perimeter of the vehicle. The camera of the system is augmented with an embedded processor based on DSP (digital signal processor) or FPGA (field-programmable gate array) technology to provide for the automatic detection of suspicious / hostile activities around the vehicle. The system is preferably provided with wireless transmitter means for alerting a person (e.g. the owner) of detected suspicious behavior.

The invention relates to an upper limb rehabilitation robot control system. An upper computer transmits instructions to an MCU (micro-programmed control unit) module through a Bluetooth / WIFI, the MCUmodule selects different training modes according to the instructions, a voice module identifies unspecific voice, voice signals are acquired and processed and transmitted to the MCU module by the aidof an SPI (serial peripheral interface) communication protocol, a driving module drives a motor to rotate in a controlled manner, so that training actions are completed, electromyography collection of a healthy side of a patient is implemented by an electromyography control module, electromyography signals are extracted, processed and analyzed, actions of the healthy side of the patient map a mechanical arm, movement intentions are identified, and functional compensation and control is achieved. By the aid of various man-machine interaction modes, targeted rehabilitation training effects canbe effectively achieved, joints of the upper limb of the patient can be accurately positioned when the system is used, and tissue damage and pulled muscle caused by misplaced training are avoided. Thecontrol system is small in size, convenient to use, good in training effect and low in manufacturing cost.