CN108108340A - For the dialogue exchange method and system of intelligent robot - Google Patents
For the dialogue exchange method and system of intelligent robot Download PDFInfo
- Publication number
- CN108108340A CN108108340A CN201711215862.6A CN201711215862A CN108108340A CN 108108340 A CN108108340 A CN 108108340A CN 201711215862 A CN201711215862 A CN 201711215862A CN 108108340 A CN108108340 A CN 108108340A
- Authority
- CN
- China
- Prior art keywords
- dialogue
- topic
- user
- interaction
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/12—Use of codes for handling textual entities
- G06F40/14—Tree-structured documents
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Human Computer Interaction (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Machine Translation (AREA)
- User Interface Of Digital Computer (AREA)
Abstract
The invention discloses a kind of dialogue interaction processing methods and system for intelligent robot, and this method comprises the following steps：During intelligent robot is interacted with user session, parsing context dialogue interactive information generates corresponding topic label, and the topic label is used for marking the topic belonging to each round dialogue interaction；The dialogue data of current round user output is obtained, user view is obtained with reference to the topic tag resolution of context dialogue interactive information；Output data is talked with according to the user view decision making.According to the present invention, after the voice messaging of user is received, with reference to current topic label, the output under same topic can be generated, ensures the continuity of dialogue, and then improves conversational quality, promotes the dialogue experience of user.
Description
Technical field
The present invention relates to field in intelligent robotics more particularly to a kind of dialogue exchange method for intelligent robot and it is
System.
Background technology
With the continuous development of science and technology, the introducing of information technology, computer technology and artificial intelligence technology, machine
Industrial circle is progressively walked out in the research of people, gradually extends to the necks such as medical treatment, health care, family, amusement and service industry
Domain.And requirement of the people for robot also conform to the principle of simplicity the multiple mechanical action of substance be promoted to anthropomorphic question and answer, independence and with
The intelligent robot that other robot interacts, human-computer interaction also just become an important factor for determining intelligent robot development.
Therefore, the interaction capabilities of intelligent robot are promoted, improves the class human nature of robot and intelligent, is the important of present urgent need to resolve
Problem.
The content of the invention
The first technical problem to be solved by the present invention is to need to provide a kind of human-computer interaction side for intelligent robot
Method and system, the program can ensure the continuity of dialogue, enhance interactive interest, improve the interactive experience of user.
In order to solve the above-mentioned technical problem, embodiments herein provides firstly a kind of dialogue for intelligent robot
Interaction processing method, this method comprises the following steps：During intelligent robot is interacted with user session, context is parsed
Talk with interactive information, generate corresponding topic label, the topic label is used for marking belonging to each round dialogue interaction
Topic；The dialogue data of current round user output is obtained, is used with reference to the topic tag resolution of context dialogue interactive information
Family is intended to；Interaction data is talked with according to the user view decision making.
Preferably, determine that model carries out topic label to every wheel dialogue and determines using topic label, the topic label
It is to carry out deep learning training by the data to more wheel dialogues under same topic to be formed to determine model.
Preferably, in the step of talking with interaction data according to the user view decision making, selected from dialog database
It selects and is intended to the dialogue interaction content of the topic tag match and the current round user session of combination, generation dialogue interaction data
And export to user, wherein, the data of the dialog database are labeled with different topic labels.
Preferably, in the dialog database, the correspondence answer mode under different topic labels is set to same problem；
After current round topic label is determined, answer mode, generation dialogue interaction data are corresponded to reference to it.
Preferably, further include：User identity is identified, judges whether active user is child user；If children
User is then based upon dialog database and dialogue label that child user builds come the interaction that engages in the dialogue.
Another aspect according to embodiments of the present invention additionally provides a kind of dialogue interaction process system for intelligent robot
System, the system include following module：Topic label determining module, during intelligent robot is interacted with user session,
Context dialogue interactive information is parsed, generates corresponding topic label, the topic label is used for marking each round dialogue friendship
Topic belonging to mutually；User view parsing module obtains the dialogue data of current round user output, talks with reference to context
The topic tag resolution of interactive information obtains user view；Dialogue data generation module is given birth to according to the user view decision-making
Into dialogue interaction data.
Preferably, the topic label determining module determines that model carries out topic to every wheel dialogue using topic label
Label determines that the topic label determines that model is to carry out deep learning instruction by the data to more wheel dialogues under same topic
Practice what is formed.
Preferably, the dialogue data generation module, from dialog database selection and pair of the topic tag match
It talks about interaction content and combines current round user session and be intended to, generation dialogue interaction data is simultaneously exported to user, wherein, it is described right
The data of words database are labeled with different topic labels.
Preferably, in the dialog database, the correspondence answer mode under different topic labels is set to same problem；
The dialogue data generation module after current round topic label is determined, corresponds to answer mode, generation dialogue is handed over reference to it
Mutual data.
Preferably, further include：User identity is identified in user identification module, whether judges active user
For child user；The dialogue data generation module when user is child user, is based upon the dialogue that child user is built
Database and topic label engage in the dialogue interaction.
Another aspect according to embodiments of the present invention additionally provides a kind of dialogue interactive system for intelligent robot,
The system includes：Cloud server possesses dialogue interaction process system as described above；Intelligent robot, acquisition and use
The multi-modal interaction data of family interaction, and the multi-modal interaction data is sent to the cloud server, it is exported to user
Dialogue alternate statement from the cloud server.
Preferably, the artificial Story machine of the intelligence machine or chat robots.
Compared with prior art, one or more of said program embodiment can have the following advantages that or beneficial to effect
Fruit：
During the embodiment of the present invention in intelligent robot with user session by interacting, parsing context dialogue interaction
Information generates corresponding topic label, then, obtains the dialogue data of current round user output, talks with reference to context and hand over
The topic tag resolution of mutual information obtains user view, and talks with output data according to the user view decision making.This hair
Bright embodiment is by the method for deep learning, training topic label generation model, in this way, talk with for arbitrary wheel, it can be true
Fixed corresponding topic label, after the voice messaging of user is received, with reference to current topic label, can generate same topic
Under output, ensure dialogue continuity, and then improve conversational quality, promoted user dialogue experience.
Other features and advantages of the present invention will be illustrated in the following description, also, partly becomes from specification
It obtains it is clear that being understood by implementing technical scheme.The purpose of the present invention and other advantages can by
Specifically noted structure and/or flow are realized and obtained in specification, claims and attached drawing.
Description of the drawings
Attached drawing is used for providing to the technical solution of the application or further understanding for the prior art, and constitution instruction
A part.Wherein, the attached drawing for expressing the embodiment of the present application is used to explain together with embodiments herein the technical side of the application
Case, but do not form the limitation to technical scheme.
Fig. 1 is the Story machine of the embodiment of the present application or the application scenarios schematic diagram of chat robots.
Fig. 2 is the Story machine of the embodiment of the present application or the functional block diagram of chat robots.
Fig. 3 is the Story machine of the embodiment of the present application or the hardware architecture diagram of chat robots.
Fig. 4 is the work(of the example one of the dialogue interaction process system of the cloud server (high in the clouds brain) of the embodiment of the present application
It can structure diagram.
Fig. 5 is the work(of the example two of the dialogue interaction process system of the cloud server (high in the clouds brain) of the embodiment of the present application
It can structure diagram.
Fig. 6 is the flow signal of the example one of the dialogue interaction processing method for intelligent robot of the embodiment of the present application
Figure.
Specific embodiment
Carry out the embodiment that the present invention will be described in detail below with reference to accompanying drawings and embodiments, how the present invention is applied whereby
Technological means solves technical problem, and the realization process for reaching relevant art effect can fully understand and implement according to this.This Shen
Each feature that please be in embodiment and embodiment, can be combined with each other under the premise of not colliding, the technical solution formed
Within protection scope of the present invention.
In addition, the flow of attached drawing can be in the computer system of such as a group of computer-executable instructions the step of illustrating
Middle execution.Also, although logical order is shown in flow charts, it in some cases, can be to be different from herein
Order performs shown or described step.
In recent years, with the rapid development of artificial intelligence, chat robots also receive the extensive of academia and industrial quarters
Concern.Chat robots are a kind of intelligent robots to be engaged in the dialogue by the natural language simulation mankind.Chat robots at present
It is broadly divided into online customer service, amusement, education, personal assistant and intelligent answer this five species.No matter which kind of robot can be with
User carries out a degree of interaction.However, in current chat robot application scene, the dialogue of robot and user are past
Toward being the question answering system in knowledge based storehouse to realize, so it is easy to cause dialogue and is always just tied quickly after several bouts
Beam, the poor expandability of chat, often so that dialogue is uninteresting and usually after two-wheeled dialogue with regard to terminating.Moreover, topic is chatted
In some topics be that some are that have the content of entity node chatting, some chat contents are the contents of no entity, in chat
In the case that appearance is no entity, it can only be supported by way of database, but for storing data in the database,
If not making topic label, also usually it is present with situation about giving an irrelevant answer.Than shown in the following example, wherein Q represents user terminal, A tables
Show robotic end：
Q：There is good-looking film to recommend
A:It hears《Roman Holiday》Well.
Q:It has seen.
A1 (mistake)：That is all right, what did the doctor say
A2 (correct)：I recommends one to you again,《Wolf is fought》.
Obviously, A1 should more be used for doing the reply of topic " sick to see the doctor " above.After the content of A1 is replied by robot,
Can very bad experience be brought to user, it is believed that robot it is intelligent relatively low, thereby reduce the product using the robot
Polarity, and correctly recommend《Wolf is fought》After film is to user, then it can preferably meet user demand.
An embodiment of the present invention provides a kind of dialogue interaction processing method and system for intelligent robot, by asking
It answers database and topic tag processes is done in often wheel dialogue interaction, suitable response, guarantee pair can be selected for same problem
The continuity of words, and then conversational quality is improved, promote the dialogue experience of user.
For a better understanding of the present invention, before being illustrated to embodiment, first to the term " words involved in this example
Topic " and " topic label " slightly illustrate.
Topic：It borrows the narrow sense in linguistics on topic to define, i.e. the central body of a word is the words
Topic.Based on it is desirable that topic can preferably extend, all central body association attributeses around topic define
Sub-topic under topic.Such as：" you like Liu De magnificent" central body is Liu Dehua, entitled " Liu De if that the words
China ", if the main topic that at this moment we define discussion is " Liu Dehua ", then " you like listening his when next sentence is asked《Ice
Rain》" central body is " Liu Dehua's《Ice rain》", then it is sub-topic.
Topic is divided into main topic and sub-topic, it is necessary to identify which is main topic, which is sub-topic, and main topic is in star
Star personage is primarily referred to as under this field, sub-topic is his direct attribute, is unfolded around star personage.In addition, such as
Fruit problem is that " you like listening《Ice rain》", the topic that the words is chatted is music, and main topic is《Ice rain》If next
Also ask that " this song is that Liu Dehua is sung, you know Liu Dehua", " Liu Dehua " that the words is chatted is then sub-topic.To sum up,
In short there may be multiple topic labels, be main topic under some topic, may be then sub-topic under another topic,
It is such to define mainly with topic extension correlation afterwards.
" topic label " is used primarily in the topic marked in human-computer dialogue interaction per round belonging to dialogue interaction.For
The question and answer sentence of each round dialogue, marks corresponding topic, then sets topic ID to the question and answer sentence for being labelled with topic and deposits
Storage is in memory.Since the dialogue output data of decision-making in this example depends on dialog database, in order to more accurately sieve
Response content is selected, topic label is advanced with and the topic of question and answer data in the dialog database of the present embodiment has also been carried out clearly
Mark.
Embodiment
Fig. 1 is the Story machine of the embodiment of the present application or the application scenarios schematic diagram of chat robots.In the application scenarios,
Including intelligent robot (also referred to as " dialogue robot ") 20 and high in the clouds brain (cloud server) 10, the intelligent robot 20 is with using
Family U carries out voice dialogue interaction.The robot 20 can also be to carry in addition to it can be tangible machine people shown in FIG. 1
Robot application program on smart machine, smart machine can be traditional PC PCs, LapTop laptops,
It line holographic projections equipment etc. or can carry and can be by wireless modes such as WLAN, mobile communications networks
Access the terminal device of internet.In the embodiment of the present application, wireless terminal includes but not limited to mobile phone, Netbook (net book)
Generally there is multi-modal information acquisition and data transmission etc. Deng, wireless terminal.High in the clouds brain 10 is used as intelligent robot 20
Brain end, be configured with dialogue interaction process system 100, which is used for handling the multimode of the transmission of intelligent robot 20
State input data, such as parse vision data, complete visual identity, vision-based detection and, perform affection computation, cognition calculate and
Semantic understanding etc. mainly talks with the voice data of user in interaction, so as to which decision-making goes out the dialogue to be output of robot 20
Voice or other multi-modal output datas.
It is pointed out that the dialogue exchange method and system of this intelligent robot, are also applied for being suitble to children's AI equipment,
Such as children-story machine (it is a kind of can meet pediatric population listen to music, children's AI equipment of story, national literature audio and video, which can
Possess the cartoon IP images of animal and personage) conversational applications scene, set in addition, the Story machine can be controlled by intelligent handhold
It is standby, it is performed with completing the setting of intelligent robot and instruction.
Illustrate the composition and work(of intelligent robot of the present invention by taking the chat robots of entity form as an example below
Energy.
Fig. 2 is the Story machine of the embodiment of the present application or the functional block diagram of chat robots.As shown in Fig. 2, the machine
People is mainly the multi-modal interaction data gathered with user interaction, and multi-modal interaction data is sent to cloud server 10,
The dialogue alternate statement from cloud server is exported to user.Robot control system mainly includes interactive information acquisition module
2110th, communication module 2120, voice-output unit 2130, robot limb control unit 2210 and attitude transducer 2220.
Interactive information acquisition module 2110 gathers outside and interactively enters information, and specific include gathers extraneous voice messaging
Voice collecting unit 2111, the touch sensor 2112 for gathering external touch pressure data and the image for gathering external image information
Collecting unit 2113.The multi-modal information that interactive information acquisition module 2110 collects is passed through networked interactive by communication module 2120
Unit 2121 is sent at high in the clouds brain 10 and is handled, and receives the friendship for carrying out 10 decision-making of high in the clouds brain goes out response user
Talk with output data or other multi-modal decision datas obtained from being mutually intended to.Networked interactive unit 2121 realizes communication module
2120 with the data interaction of high in the clouds brain 10.Voice-output unit 2130 exports matched voice according to speech-controlled information and returns
It should.Robot limb control unit 2210 exports matched robot limb control signal with driving machine according to action control information
The limbs of device people make corresponding actions.2220 supervisory-controlled robot current pose of attitude transducer, thus can be to avoid robot
Ignore the current posture of itself and perform action by force, hair situations such as so as to which movement posture mistake or disequilibrium be avoided to fall down
It is raw.
In view of the power supply requirement of each function module, data processing needs and difference functionally, the chat machine
The electronic control system of people is configured to master system and lower computer system two parts.Master system and slave computer system
It unites and respectively includes one piece of independent master control borad respectively, the external circuit original paper of master system and lower computer system is connected to each
Master control borad on.So the module that resource has conflict is separated on the premise of the integrated level of guarantee system entirety, so as to ensure
The stable and high effective operation of system.
In this example, different function modules is distributed to master system and slave computer system according to mode shown in Fig. 2
In system.Specifically, robot limb control unit 2210 and attitude transducer 2220 are constructed in lower computer system 220
In, other function modules are constructed in master system 210.
In the present embodiment, system can also include the electricity quantity display module of the current information about power of display robot.It examines
Consider electricity and show that the data processing amount needed is not high but certain power drives is needed to support (driving light emitting diode), because
This electricity quantity display module is arranged in lower computer system.In concrete operations, the master control borad acquisition of master system and transmitter
The current information about power of device people, electricity quantity display module export corresponding electricity according to information about power and show.
Further, understand the current interaction mode of robot for the ease of user, be also provided in lower computer system
Show the interactive display module of the current interaction mode of robot.In concrete operations, the master control borad acquisition of master system is simultaneously
The current interaction mode of distribution of machine people, interaction mode include recording state, voice/action output state and semantic parsing shape
State；Interactive display module exports corresponding interaction mode according to interaction mode and shows.
Hardware block diagram as shown in Figure 3, the master control borad of master system is the master based on full will double-core A20 processors
Plate is controlled, and is integrated with wireless networking module (WiFi), Mike's noise reduction module and audio amplification block.Wherein, A20 processors into
The external Preprocessing for interactively entering information of row can generate the action directive of robot movement；WiFi networking modules
Realize the data interaction with high in the clouds brain 10；Mike's noise reduction module and the microphone being connected on master control borad realize extraneous voice
The acquisition of information；Audio amplification block and the loud speaker being connected on master control borad realize the output of speech response.
The interface that host computer master control borad 210 provides has：Capacitance touch interface 212, three-wire interface, line sequence are power supply (VCC)
(GND) output (OUT) is grounded, is connected to touch modules 204；Serial communication interface 216, three-wire interface, line sequence are ground connection
(GND) uplink (RX) downlink (TX) is connected to the serial communication interface 217 of slave computer master control borad 220；Speaker interface 213,
Two line interfaces, line sequence bears (Speaker-) for audio signal just (Speaker+) audio signal (in this example, to be had 2 to raise one's voice
Device interface), it is connected to loud speaker 205；Microphone interface 211, two line interfaces, line sequence are the microphone signal is positive (Mic+) microphone
Signal bears (Mic-), is connected to microphone 203；Charge port 214, two line interfaces, line sequence are grounded (GND) for power supply (VCC),
It is connected to robot charge port 201 and connects power management module 215；The battery charging interface of power management module 215, two lines
Interface, line sequence are grounded (GND) for power input (DCIN), are connected to lithium battery 202.
The master control borad of lower computer system is the master control borad based on STMicw Electronics microcontroller STM32, is integrated with six thereon
Axis attitude transducer MPU6500 and motor drive module.Wherein, microcontroller STM32 generates robot limb control signal；
Six axis attitude transducer MPU6500 supervisory-controlled robot current poses；Motor drive module driving robot limb action.
The interface that slave computer master control borad 220 provides has：Power interface, two line interfaces, line sequence are VCC GND (not shown),
The voltage stabilizing chip 223 of power management module passes through power interface and the power management of lithium battery 202 and host computer master control borad 210
Module 215 is connected；Serial communication interface 227, three-wire interface, line sequence are GND RX TX, host computer master control borad 210 and bottom
Data transmission is realized by serial communication between machine master control borad 220；Three road motor interfaces, are two line interfaces per road, and line sequence is electricity
Machine just (Motor+) motor is born (Motor-), and motor drives (224,225,226) to drive robot by three road motor interfaces
Motor (231,232,233) (three motors are respectively two thigh motors, one arm motor) operates to realize that robot moves
Make；System charge display interface, four-wire interface, line sequence export (IO) for output (IO) and export (IO) ground connection (GND), are connected to electricity
Measure display lamp 206 (electricity display lamp is multi-color LED (LED) lamp, and three I/O interfaces correspond to red R, green G, indigo plant B respectively)；
Interactive display interface, two line interfaces, line sequence are PWM GND, and being connected to interactive display lamp 207, (interaction display lamp is respiratory nasal
Lamp).
System upper and lower computer master control borad 210 and 220 is controlled by a physical switch.System boot process is：
System electrification, host computer master control borad 210 complete networking, initialization；
The interaction display lamp 207 of slave computer master control borad 220 is in breathing state and host computer master control borad 210 is waited to initialize
Into.
Interaction：
The initialization of host computer master control borad 210 is completed, and upper and lower computer master control borad 210 and 220 is normal by serial ports (216,217)
Transfer data；
Microphone 203 gathers audio signal, and the processing of host computer master control borad 210 is given by the noise reduction to audio and amplification
Voice messaging is passed to high in the clouds brain 10 by chip A20, A20 by networking module, and high in the clouds brain 10 returns to multi-modal decision-making number
A20 is passed back according to through networking module, and A20 controlling loudspeakers 205 feed back to user speech response, and at the same time A20 is by action control
Information (needing the action performed), information about power and interaction mode information are sent to slave computer master control borad 220 by serial ports；
Slave computer master control borad 220 by serial ports receive A20 control instruction, complete electricity show, interaction show, leg and
The multi-modal interactive action such as hand motion.
During electricity is shown, system charge is showed by RGB tri coloured lanterns：R represents not enough power supply, and B is representing electricity just
Often, G represents electricity abundance.Host computer master control borad 210 informs user's charge condition by loud speaker simultaneously.In interaction display process
In, pass through the LED light of PWM controls：Being always on prompting user, robot is in recording state at this time, and often go out prompting user machine at this time
People's voice output, flash for prompting user robot carry out semantic parsing in networking.
The each component and function of the dialogue interactive system 100 of high in the clouds brain 10 are illustrated below.
As shown in figure 4, dialogue interactive system 100 includes topic label determining module 110, user view parsing module 120
With dialogue data generation module 130.The function of above-mentioned modules is specifically described below.
Topic label determining module 110 during intelligent robot is interacted with user session, parses context pair
Interactive information is talked about, generates corresponding topic label, which is used for marking the topic belonging to each round dialogue interaction.
Specifically, topic label determining module 110, the voice messaging after receiving communication module 2120 and forwarding
Afterwards, respond the voice messaging and generate corresponding text message.First, language is carried out to the voice messaging after such as noise suppression preprocessing
The comprehensive analysis of sound identification, generates text message corresponding with voice messaging.Then, text analyzing is carried out to text message, i.e.,
Obtain the specific semantic content of text.Specifically, after recognition result is obtained, identification is tied using natural language processing technique
Fruit carries out semantic parsing.Semantic analysis refers to for given natural language being converted into certain formalization table for reflecting its meaning
Show, that is, by the mankind it will be appreciated that natural language be converted into computer it will be appreciated that formal language.Obtaining parsing knot
After fruit, the semantic similarity (similarity of problem and problem) of the content in the analysis result and the knowledge base that has set is calculated,
So as to the data that search matches with analysis result in knowledge base.So far, complete and the parsing for talking with interactive information is operated.
It, can be by judging to whether there is in obtained speech text information and the relevant spy of topic after semantic understanding
Vocabulary is determined to determine the topic of text information." specific vocabulary " is the prior vocabulary relevant with topic or short being set
Language, for example, the contents such as the name of star, name of film.Moreover, those skilled in the art can use according to current web technology
Language or user demand update or addition " specific vocabulary ", the content for making database is more abundant, improves user experience.It can travel through
Obtained speech text information and each specific vocabulary are carried out morphology similarity by each vocabulary in " specific vocabulary " database
And/or Semantic Similarity Measurement, judge to whether there is corresponding specific vocabulary in speech text information.When morphology similarity is more than
Threshold value, and numerical value is very big, then can determine whether that speech text there are specific vocabulary, otherwise, calculates language without computing semantic similarity
The weighted sum of adopted similarity and morphology similarity is to determine whether there are specific vocabulary.The speech text information obtained on judgement
In whether there is specific vocabulary method, can also be realized by other technologies, do not limited herein.
If not finding specific vocabulary, anticipated according to the topic of former wheel dialogues to parse the user of dialogue interactive information
Figure, and topic is determined based on user view.Shown in following example：
Q：Nearest action movie《Warwolf 2》It sees very well, you have seen
A：I does not see.
When the A contents in talking with interactive information to the wheel carry out topic judgement, due to not finding matched specific word
It converges, then can not merely be remitted by specific word and determine its subject information, therefore, with reference to theme-film of previous alternate statement
《Warwolf 2》, then the user view that can determine A contents is " not see and flash back past events《Warwolf 2》", it is possible thereby to determine that topic is still
Film《Warwolf 2》.The correspondence of dialogue alternate statement of the topic label with extracting the topic label is then preserved a certain
In memory, such as：
Q：Nearest action movie《Warwolf 2》It sees very well, you have seen【Theme-film《Warwolf 2》】
A：I does not see.【Theme-film《Warwolf 2》】
During the topic content of next round dialogue interactive information to be determined, handed over by transferring context dialogue from the memory
Mutual theme can just complete topic and determine to handle well.
It is determined except through searching the mode of specific vocabulary beyond topic label, in a preferable example, topic mark
Determining module 110 is signed, determines that model carries out topic label to every wheel dialogue and determines using topic label, topic mark therein
Label determine that model is to carry out deep learning training by the data to more wheel dialogues under same topic to be formed.
Specific learning method is as follows：
Step 1, the sample information for being trained to default grader is obtained.In the present embodiment, multiple words are selected
More wheel dialogue datas under topic, the sample dialogue data being utilized respectively under each topic are trained default grader.It is excellent
Selection of land can gather and be used as sample data labeled as the voice of different themes by manual sort in history, before being trained
Voice messaging is converted into Text Mode.
Step 2, sample data is pre-processed, the noises text such as removal such as " eh, " therein is instructed
Practice text.
Step 3, the text feature of training text is extracted.
Specifically, cutting word processing can be carried out according to default step-length to training text, it is special to obtain text based on cutting word result
Sign.
Step 4, the text feature of training text is input in grader and be trained, obtain object classifiers.
After topic label determining module 110 is completed to turn text-processing to the voice of the often wheel dialogue interactive information of input,
Obtained text input will be handled into object classifiers, you can obtain the theme of round dialogue interactive information.Pass through we
Method can be in the case of the content incorporeity information that speech-to-text is converted to, for example, dialogue includes the nothing of " without "
It during the content of entity information, remains to accurately determine its subject content, and the upper specific vocabulary mentioned is looked into processing speed
Inquiry method is more quick.
User view parsing module 120 obtains the dialogue data of current round user output, talks with reference to context and hand over
The topic tag resolution of mutual information obtains user view.
For information content compared in the user session data of horn of plenty, for example, the content including entity information, user view solution
Analysis module 120 can be operated according to the semantic understanding of topic label determining module 110, and voice messaging is generated corresponding text envelope
Then breath carries out semantic understanding and obtains user view.It is more succinct in view of certain user's conversation content, do not include generally having
The content of practical significance, such as the content of these imperfect semantemes such as " not having ", " having seen ", " not seeing also ", can not simple root
Carry out the true intention of Direct Recognition user according to the corresponding text of voice, therefore, when parsing user view, preferably in combination with context
Talk with the topic label of interaction to be identified.
With reference to above example, the analysis result obtained by speech recognition is " I does not see ", and a upper conversation content
Subject information is：Film-《Warwolf 2》, therefore, the true intention of this dialogue interaction is can determine by the way that the two is combined is
" there is no film《Warwolf 2》”.For compared with prior art, if merely obtained according to the analysis result of current session content
User view can be varied, then easilying lead to the voice messaging replied, there are larger deviation, bands with actual intention
Carry out bad user experience, and this example by combine the subject information that context is talked with determine to be intended to then can preferably solve it is above-mentioned
Problem.
Dialogue data generation module 130 talks with interaction data according to user view decision making.
Specifically, dialogue data generation module 130, from the selection of dialog database 140 and the dialogue of topic tag match
Interaction content simultaneously combines current round user session intention, and generation dialogue interaction data is simultaneously exported to user, wherein, dialogue data
The data in storehouse 140 are labeled with different topic labels.
In dialog database 140, the correspondence answer mode under different topic labels is set to same problem.It is specific next
Say, in the database 140, store the list of problem and response content, structure be by same problem may (probability compared with
Greatly) corresponding response mode arranges out, and carries out topic mark for the problem of imperfect semanteme according to response content.Such as
Lower shown list：
Wherein specific topic label determines method, and topic label may be employed and determine model to realize, will reply interior
Hold input into the model, and then obtain corresponding label.By stamping topic label to the data in database so that work as user
The problem of input, Q in multiple topics when be likely to occur, and can select appropriate reply.
After current round topic label is determined, answer mode is corresponded to reference to it for dialogue data generation module 130, raw
Into dialogue interaction data.There is matched output masterplate for the response mode set in database, given birth to based on the output masterplate
Into dialogue interaction data.
Topic 1：
Q：There is good-looking film to recommend
A:It hears《Roman Holiday》Well.
Q:It has seen.
A2 (correct)：I recommends one to you again, has just shown《Wolf war 2》It is well received.
Topic 2：
Q：I catches a cold.
A:Pitiful is well-behaved, goes to have seen doctor
Q:It has seen.
A2：That is all right, what did the doctor say
In addition, in other examples, as shown in figure 5, the dialogue interaction process system 100 of the present invention can also include user
Identification module 150, is identified user identity, judges whether active user is child user；If child user,
It is based upon dialog database that child user builds and topic label engages in the dialogue interaction.
It specifically may be referred to following child user recognition methods.For example, image acquisition units 2113 gather active user
Face information, the user identification module 150 of high in the clouds brain 10 is sent to by communication module 2120, elder generation is from scene
It detects the presence of face and determines its position.Then, after face is detected, recognition of face is carried out, will be had detected that
Face to be identified matches compared with the different type face in database, obtains relevant information.Recognition of face can adopt
The method for taking the method and template matches of extraction Face geometric eigenvector, the method that template matches are preferentially taken in this example.In addition, also
It can identify whether active user is child user by way of sound characteristic detection, such as to voice that user is inputted
It is identified, judges whether voice is child's voice.In the present embodiment, voice is provided in user identification module 150 in advance
Identification model, the voice that can be inputted by speech recognition modeling to user are identified, to determine the classification of the voice.It should
Speech recognition modeling can be machine learning model, after the training and study to great amount of samples data, the machine learning mould
Type can classify to the classification of voice., it is necessary to be trained to grader before testing voice, to get
Object classifiers.Specifically include following steps：
Step 1, the sample voice for being trained to default grader is obtained.It, can be to children in the present embodiment
Sound carry out sampling as sample voice, default grader is trained using the sample voice that these are collected.It is excellent
Selection of land can gather and be used as sample voice labeled as the voice of child's voice by manual sort in history.
Step 2, to sample voice carry out voice activity detection removal training data in it is mute, obtain train voice.
Step 3, the acoustic feature of training voice is extracted.
Specifically, framing can be carried out according to default step-length to training voice, then according to default step-length to training voice
Every frame extraction acoustic feature, wherein, acoustic feature can be wave filter group (Filter bank40, abbreviation Fbank40) feature
Or mel-frequency cepstrum coefficient (Mel-Frequency Cepstral Coefficients, abbreviation MFCC) feature.
Step 4, the acoustic feature of training voice is input in grader and be trained, obtain object classifiers.
User identification module 150 is set to be mainly in view of：It is more various in dialog database in high in the clouds brain 10,
Sensitive content or the bad network information may be included.User identification module 150 is after child user is identified, cloud
Interaction that end brain 10 is further selected as the dialog database that child user builds and dialogue label engages in the dialogue, can so keep away
Exempt to send sensitive content to the child user, prevent that the physical and mental health to children from adversely affecting.
The structure for the dialog database built for child user is similar with foregoing database structure, but question and answer therein
Content mainly includes the intelligence developments class contents such as children education, amusement, and some sensitive, hard to understand information are then shielded, base
In this kind of dialog database, the dialogue suitable for children can be targetedly exported, and similar to this sensitivity of adult entertainment's culture
Property content will not then be pushed in human-computer interaction.
Fig. 6 is the flow diagram of the example one of the dialogue exchange method for intelligent robot of the embodiment of the present application.
Illustrate the interaction flow of this interactive system with reference to Fig. 6.
As shown in fig. 6, in step S610, topic label determining module 110 is interacted in intelligent robot with user session
During, parsing context dialogue interactive information generates corresponding topic label, and the topic label is used for marking each round
Topic belonging to secondary dialogue interaction；In step S620, user view parsing module 120 obtains pair of current round user output
Data are talked about, user view is obtained with reference to the topic tag resolution of context dialogue interactive information；In step s 130, dialogue data
Generation module 130 talks with interaction data according to the user view decision making.
In embodiments of the present invention, by the method for deep learning, training topic label generation model is appointed in this way, being directed to
Meaning wheel dialogue, can determine corresponding topic label, after the voice messaging of user is received, with reference to current topic mark
Label, can generate the output under same topic, ensure the continuity of dialogue, and then improve conversational quality, promote the dialogue of user
Experience.
Supplementary notes
During the artificial Story machine of intelligence machine in the present embodiment, in addition to the characteristics of being described above, further may be used also
To possess following feature.
(1) Story machine with wechat can be realized and interconnect as a part for family's Internet of Things；
(2) possesses program request, the functions such as collection, voice control interrupt, sound；
(3) possess OCR (Optical Character Recognition, optical character identification) function, realize that voice is read
Paint the function of this and reading；
(4) can content push actively be carried out according to the hobby of child user.
Since the method for the present invention describes what is realized in computer systems.The computer system can for example be set
In control core processor.For example, method described herein can be implemented as that software can be performed with control logic, by
CPU in operating system is performed.Function as described herein can be implemented as being stored in readable Jie of non-transitory tangible computer
Program instruction set in matter.When implemented in this fashion, which includes one group of instruction, when the group is instructed by counting
It promotes computer to perform the method that can implement above-mentioned function when calculation machine is run.Programmable logic can be installed temporarily or permanently
In non-transitory visible computer readable medium, such as ROM chip, computer storage, disk or other storages
Medium.Except with software come in addition to realizing, logic as described herein can utilize discrete parts, integrated circuit and programmable logic
Equipment (such as, field programmable gate array (FPGA) or microprocessor) be used in combination programmable logic or including them
Any other equipment of any combination embodies.All such embodiments are intended to fall under within the scope of the present invention.
It should be understood that disclosed embodiment of this invention is not limited to processing step disclosed herein, and should prolong
Reach the equivalent substitute for these features that those of ordinary skill in the related art are understood.It is to be further understood that it uses herein
Term be used only for the purpose of describing specific embodiments, and be not intended to limit.
" one embodiment " or " embodiment " mentioned in specification means the special characteristic described in conjunction with the embodiments, structure
Or characteristic is included at least one embodiment of the present invention.Therefore, the phrase " reality that specification various places throughout occurs
Apply example " or " embodiment " same embodiment might not be referred both to.
While it is disclosed that embodiment as above, but the content only to facilitate understand the present invention and adopt
Embodiment is not limited to the present invention.Any those skilled in the art to which this invention pertains are not departing from this
On the premise of the disclosed spirit and scope of invention, any modification and change can be made in the implementing form and in details,
But the scope of patent protection of the present invention, still should be subject to the scope of the claims as defined in the appended claims.
Claims (12)
1. a kind of dialogue interaction processing method for intelligent robot, this method comprises the following steps：
During intelligent robot is interacted with user session, parsing context dialogue interactive information generates corresponding topic
Label, the topic label are used for marking the topic belonging to each round dialogue interaction；
The dialogue data of current round user output is obtained, is used with reference to the topic tag resolution of context dialogue interactive information
Family is intended to；
Interaction data is talked with according to the user view decision making.
2. according to the method described in claim 1, it is characterized in that,
Determine that model carries out topic label to every wheel dialogue and determines using topic label,
The topic label determines that model is to carry out deep learning by the data to more wheel dialogues under same topic to train come shape
Into.
3. method according to claim 1 or 2, which is characterized in that handed over being talked with according to the user view decision making
In the step of mutual data,
It anticipates from dialog database selection and the dialogue interaction content and the current round user session of combination of the topic tag match
Figure, generation dialogue interaction data are simultaneously exported to user, wherein, the data of the dialog database are labeled with different topic marks
Label.
4. according to the method described in claim 3, it is characterized in that,
In the dialog database, the correspondence answer mode under different topic labels is set to same problem；
After current round topic label is determined, answer mode, generation dialogue interaction data are corresponded to reference to it.
5. according to Claims 1 to 4 any one of them method, which is characterized in that further include：
User identity is identified, judges whether active user is child user；
If child user, then it is based upon dialog database and dialogue label that child user builds come the interaction that engages in the dialogue.
6. a kind of dialogue interaction process system for intelligent robot, which includes following module：
Topic label determining module, during intelligent robot is interacted with user session, parsing context dialogue interaction
Information, generates corresponding topic label, and the topic label is used for marking the topic belonging to each round dialogue interaction；
User view parsing module obtains the dialogue data of current round user output, talks with interactive information with reference to context
Topic tag resolution obtain user view；
Dialogue data generation module talks with interaction data according to the user view decision making.
7. system according to claim 6, which is characterized in that
The topic label determining module determines that model carries out topic label to every wheel dialogue and determines using topic label,
The topic label determines that model is to carry out deep learning by the data to more wheel dialogues under same topic to train come shape
Into.
8. the system according to claim 6 or 7, which is characterized in that
The dialogue data generation module, from the dialogue interaction content of dialog database selection and the topic tag match simultaneously
It is intended to reference to current round user session, generation dialogue interaction data is simultaneously exported to user, wherein, the number of the dialog database
According to being labeled with different topic labels.
9. system according to claim 8, which is characterized in that
In the dialog database, the correspondence answer mode under different topic labels is set to same problem；
After current round topic label is determined, answer mode, generation pair are corresponded to reference to it for the dialogue data generation module
Talk about interaction data.
10. according to claim 6~9 any one of them system, which is characterized in that further include：
User identity is identified in user identification module, judges whether active user is child user；
The dialogue data generation module when user is child user, is based upon the dialog database that child user is built
It engages in the dialogue interaction with topic label.
11. a kind of dialogue interactive system for intelligent robot, which includes：
Cloud server possesses such as the dialogue interaction process system any one of claim 6~10；
Intelligent robot, acquisition and the multi-modal interaction data of user interaction, and the multi-modal interaction data is sent to
The cloud server exports the dialogue alternate statement from the cloud server to user.
12. dialogue interactive system according to claim 11, which is characterized in that
The artificial Story machine of the intelligence machine or chat robots.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711215862.6A CN108108340B (en) | 2017-11-28 | 2017-11-28 | Dialogue interaction method and system for intelligent robot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711215862.6A CN108108340B (en) | 2017-11-28 | 2017-11-28 | Dialogue interaction method and system for intelligent robot |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108108340A true CN108108340A (en) | 2018-06-01 |
| CN108108340B CN108108340B (en) | 2021-07-23 |
Family
ID=62208505
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711215862.6A Active CN108108340B (en) | 2017-11-28 | 2017-11-28 | Dialogue interaction method and system for intelligent robot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108108340B (en) |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108897723A (en) * | 2018-06-29 | 2018-11-27 | 北京百度网讯科技有限公司 | The recognition methods of scene dialog text, device and terminal |
| CN108900612A (en) * | 2018-06-29 | 2018-11-27 | 百度在线网络技术（北京）有限公司 | Method and apparatus for pushed information |
| CN109005276A (en) * | 2018-09-21 | 2018-12-14 | 马鑫 | Intelligent dialogue robot, dialog control method, computer equipment and storage medium |
| CN109086351A (en) * | 2018-07-17 | 2018-12-25 | 北京光年无限科技有限公司 | A kind of method and user tag system obtaining user tag |
| CN109359211A (en) * | 2018-11-13 | 2019-02-19 | 平安科技（深圳）有限公司 | Data-updating method, device, computer equipment and the storage medium of interactive voice |
| CN109542389A (en) * | 2018-11-19 | 2019-03-29 | 北京光年无限科技有限公司 | Sound effect control method and system for the output of multi-modal story content |
| CN109710941A (en) * | 2018-12-29 | 2019-05-03 | 上海点融信息科技有限责任公司 | User's intension recognizing method and device based on artificial intelligence |
| CN109739962A (en) * | 2018-12-26 | 2019-05-10 | 广州灵聚信息科技有限公司 | A kind of control method and device of Chat mode |
| CN109902834A (en) * | 2019-01-28 | 2019-06-18 | 北京怡凯智能技术有限公司 | A kind of old man's company active interlocution robot of topic driving |
| CN110059169A (en) * | 2019-01-25 | 2019-07-26 | 邵勃 | Intelligent robot chat context realization method and system based on corpus labeling |
| CN110297617A (en) * | 2019-06-28 | 2019-10-01 | 北京蓦然认知科技有限公司 | A kind of initiating method and device of active interlocution |
| WO2019233219A1 (en) * | 2018-06-07 | 2019-12-12 | 腾讯科技（深圳）有限公司 | Dialogue state determining method and device, dialogue system, computer device, and storage medium |
| CN110765312A (en) * | 2018-07-10 | 2020-02-07 | 阿里巴巴集团控股有限公司 | Man-machine interaction and content search method, device, equipment and storage medium |
| CN110995569A (en) * | 2019-11-12 | 2020-04-10 | 腾讯科技（深圳）有限公司 | Intelligent interaction method and device, computer equipment and storage medium |
| CN111737408A (en) * | 2019-03-25 | 2020-10-02 | 阿里巴巴集团控股有限公司 | Dialogue method and equipment based on script and electronic equipment |
| CN112100368A (en) * | 2020-07-21 | 2020-12-18 | 深思考人工智能科技(上海)有限公司 | Recognition method and device for conversation interaction intention |
| WO2021082133A1 (en) * | 2019-10-28 | 2021-05-06 | 苏州思必驰信息科技有限公司 | Method for switching between man-machine dialogue modes |
| WO2021139243A1 (en) * | 2020-01-08 | 2021-07-15 | 平安科技（深圳）有限公司 | Data processing method, device, and apparatus employing human-computer interaction, and storage medium |
| CN115905490A (en) * | 2022-11-25 | 2023-04-04 | 北京百度网讯科技有限公司 | Man-machine interaction dialogue method, device and equipment |
| CN116383365A (en) * | 2023-06-01 | 2023-07-04 | 广州里工实业有限公司 | Learning material generation method and system based on intelligent manufacturing and electronic equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100174523A1 (en) * | 2009-01-06 | 2010-07-08 | Samsung Electronics Co., Ltd. | Multilingual dialogue system and controlling method thereof |
| CN105068661A (en) * | 2015-09-07 | 2015-11-18 | 百度在线网络技术（北京）有限公司 | Man-machine interaction method and system based on artificial intelligence |
| CN105138671A (en) * | 2015-09-07 | 2015-12-09 | 百度在线网络技术（北京）有限公司 | Human-computer interaction guiding method and device based on artificial intelligence |
| CN106991123A (en) * | 2017-02-27 | 2017-07-28 | 北京光年无限科技有限公司 | A kind of man-machine interaction method and device towards intelligent robot |
| CN107193978A (en) * | 2017-05-26 | 2017-09-22 | 武汉泰迪智慧科技有限公司 | A kind of many wheel automatic chatting dialogue methods and system based on deep learning |
-
2017
- 2017-11-28 CN CN201711215862.6A patent/CN108108340B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100174523A1 (en) * | 2009-01-06 | 2010-07-08 | Samsung Electronics Co., Ltd. | Multilingual dialogue system and controlling method thereof |
| CN105068661A (en) * | 2015-09-07 | 2015-11-18 | 百度在线网络技术（北京）有限公司 | Man-machine interaction method and system based on artificial intelligence |
| CN105138671A (en) * | 2015-09-07 | 2015-12-09 | 百度在线网络技术（北京）有限公司 | Human-computer interaction guiding method and device based on artificial intelligence |
| CN106991123A (en) * | 2017-02-27 | 2017-07-28 | 北京光年无限科技有限公司 | A kind of man-machine interaction method and device towards intelligent robot |
| CN107193978A (en) * | 2017-05-26 | 2017-09-22 | 武汉泰迪智慧科技有限公司 | A kind of many wheel automatic chatting dialogue methods and system based on deep learning |
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019233219A1 (en) * | 2018-06-07 | 2019-12-12 | 腾讯科技（深圳）有限公司 | Dialogue state determining method and device, dialogue system, computer device, and storage medium |
| CN108900612A (en) * | 2018-06-29 | 2018-11-27 | 百度在线网络技术（北京）有限公司 | Method and apparatus for pushed information |
| CN108897723A (en) * | 2018-06-29 | 2018-11-27 | 北京百度网讯科技有限公司 | The recognition methods of scene dialog text, device and terminal |
| CN110765312A (en) * | 2018-07-10 | 2020-02-07 | 阿里巴巴集团控股有限公司 | Man-machine interaction and content search method, device, equipment and storage medium |
| CN109086351A (en) * | 2018-07-17 | 2018-12-25 | 北京光年无限科技有限公司 | A kind of method and user tag system obtaining user tag |
| CN109086351B (en) * | 2018-07-17 | 2022-03-22 | 北京光年无限科技有限公司 | Method for acquiring user tag and user tag system |
| CN109005276A (en) * | 2018-09-21 | 2018-12-14 | 马鑫 | Intelligent dialogue robot, dialog control method, computer equipment and storage medium |
| CN109359211A (en) * | 2018-11-13 | 2019-02-19 | 平安科技（深圳）有限公司 | Data-updating method, device, computer equipment and the storage medium of interactive voice |
| CN109359211B (en) * | 2018-11-13 | 2024-05-07 | 平安科技（深圳）有限公司 | Voice interaction data updating method, device, computer equipment and storage medium |
| CN109542389A (en) * | 2018-11-19 | 2019-03-29 | 北京光年无限科技有限公司 | Sound effect control method and system for the output of multi-modal story content |
| CN109739962A (en) * | 2018-12-26 | 2019-05-10 | 广州灵聚信息科技有限公司 | A kind of control method and device of Chat mode |
| CN109710941A (en) * | 2018-12-29 | 2019-05-03 | 上海点融信息科技有限责任公司 | User's intension recognizing method and device based on artificial intelligence |
| CN110059169A (en) * | 2019-01-25 | 2019-07-26 | 邵勃 | Intelligent robot chat context realization method and system based on corpus labeling |
| CN110059169B (en) * | 2019-01-25 | 2023-12-01 | 邵勃 | Intelligent robot chat context implementation method and system based on corpus labeling |
| CN109902834A (en) * | 2019-01-28 | 2019-06-18 | 北京怡凯智能技术有限公司 | A kind of old man's company active interlocution robot of topic driving |
| CN111737408A (en) * | 2019-03-25 | 2020-10-02 | 阿里巴巴集团控股有限公司 | Dialogue method and equipment based on script and electronic equipment |
| CN111737408B (en) * | 2019-03-25 | 2024-05-03 | 阿里巴巴集团控股有限公司 | Dialogue method and equipment based on script, and electronic equipment |
| CN110297617B (en) * | 2019-06-28 | 2021-05-14 | 北京蓦然认知科技有限公司 | Method and device for initiating active conversation |
| CN110297617A (en) * | 2019-06-28 | 2019-10-01 | 北京蓦然认知科技有限公司 | A kind of initiating method and device of active interlocution |
| WO2021082133A1 (en) * | 2019-10-28 | 2021-05-06 | 苏州思必驰信息科技有限公司 | Method for switching between man-machine dialogue modes |
| CN110995569A (en) * | 2019-11-12 | 2020-04-10 | 腾讯科技（深圳）有限公司 | Intelligent interaction method and device, computer equipment and storage medium |
| WO2021139243A1 (en) * | 2020-01-08 | 2021-07-15 | 平安科技（深圳）有限公司 | Data processing method, device, and apparatus employing human-computer interaction, and storage medium |
| CN112100368A (en) * | 2020-07-21 | 2020-12-18 | 深思考人工智能科技(上海)有限公司 | Recognition method and device for conversation interaction intention |
| CN112100368B (en) * | 2020-07-21 | 2024-01-26 | 深思考人工智能科技(上海)有限公司 | Method and device for identifying dialogue interaction intention |
| CN115905490A (en) * | 2022-11-25 | 2023-04-04 | 北京百度网讯科技有限公司 | Man-machine interaction dialogue method, device and equipment |
| CN115905490B (en) * | 2022-11-25 | 2024-03-22 | 北京百度网讯科技有限公司 | Man-machine interaction dialogue method, device and equipment |
| CN116383365A (en) * | 2023-06-01 | 2023-07-04 | 广州里工实业有限公司 | Learning material generation method and system based on intelligent manufacturing and electronic equipment |
| CN116383365B (en) * | 2023-06-01 | 2023-09-08 | 广州里工实业有限公司 | Learning material generation method and system based on intelligent manufacturing and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108108340B (en) | 2021-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108108340A (en) | For the dialogue exchange method and system of intelligent robot | |
| CN108000526A (en) | Dialogue exchange method and system for intelligent robot | |
| JP6816925B2 (en) | Data processing method and equipment for childcare robots | |
| CN110288978B (en) | Speech recognition model training method and device | |
| CN108536681B (en) | Intelligent question-answering method, device, equipment and storage medium based on emotion analysis | |
| US10963493B1 (en) | Interactive game with robot system | |
| CN103035135B (en) | Children cognitive system based on augment reality technology and cognitive method | |
| CN107870994A (en) | Man-machine interaction method and system for intelligent robot | |
| CN107704612A (en) | Dialogue exchange method and system for intelligent robot | |
| CN106297773B (en) | A kind of neural network acoustic training model method | |
| CN110444191A (en) | A kind of method, the method and device of model training of prosody hierarchy mark | |
| CN106997243B (en) | Speech scene monitoring method and device based on intelligent robot | |
| CN107199572A (en) | A kind of robot system and method based on intelligent auditory localization and Voice command | |
| WO2019062931A1 (en) | Image processing apparatus and method | |
| CN112204564A (en) | System and method for speech understanding via integrated audio and visual based speech recognition | |
| CN110490213A (en) | Image-recognizing method, device and storage medium | |
| CN110534099A (en) | Voice wakes up processing method, device, storage medium and electronic equipment | |
| CN109522835A (en) | Children's book based on intelligent robot is read and exchange method and system | |
| CN110347792A (en) | Talk with generation method and device, storage medium, electronic equipment | |
| CN110570840B (en) | Intelligent device awakening method and device based on artificial intelligence | |
| CN109102802A (en) | System for handling user spoken utterances | |
| CN106486122A (en) | A kind of intelligent sound interacts robot | |
| CN110444210A (en) | A kind of method of speech recognition, the method and device for waking up word detection | |
| CN110248021A (en) | A kind of smart machine method for controlling volume and system | |
| CN112735418A (en) | Voice interaction processing method and device, terminal and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |