CN107945619A - Learn the method, apparatus and learning robot of language - Google Patents
Learn the method, apparatus and learning robot of language Download PDFInfo
- Publication number
- CN107945619A CN107945619A CN201711132285.4A CN201711132285A CN107945619A CN 107945619 A CN107945619 A CN 107945619A CN 201711132285 A CN201711132285 A CN 201711132285A CN 107945619 A CN107945619 A CN 107945619A
- Authority
- CN
- China
- Prior art keywords
- study
- learning
- learner
- archives
- language
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 59
- 230000008569 process Effects 0.000 claims abstract description 30
- 230000008878 coupling Effects 0.000 claims abstract description 16
- 238000010168 coupling process Methods 0.000 claims abstract description 16
- 238000005859 coupling reaction Methods 0.000 claims abstract description 16
- 230000001755 vocal effect Effects 0.000 claims description 20
- 238000012549 training Methods 0.000 abstract description 33
- 230000004069 differentiation Effects 0.000 abstract description 11
- 238000010586 diagram Methods 0.000 description 14
- 238000005457 optimization Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 6
- 230000008859 change Effects 0.000 description 4
- 238000004891 communication Methods 0.000 description 4
- 239000000284 extract Substances 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000001149 cognitive effect Effects 0.000 description 2
- 230000000052 comparative effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000005021 gait Effects 0.000 description 2
- 238000012552 review Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000007257 malfunction Effects 0.000 description 1
- 238000000465 moulding Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
- G09B—EDUCATIONAL OR DEMONSTRATION APPLIANCES; APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND, DEAF OR MUTE; MODELS; PLANETARIA; GLOBES; MAPS; DIAGRAMS
- G09B19/00—Teaching not covered by other main groups of this subclass
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/20—Education
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- Educational Administration (AREA)
- Educational Technology (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Electrically Operated Instructional Devices (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Present invention is disclosed a kind of method, apparatus and learning robot for learning language, wherein, learn the method for language, including：The biocompatibility characteristics of identification learning person；Judge whether the study archives associated with the biocompatibility characteristics；If in the presence of according to the study archives coupling learning course.Different people is identified according to the difference of the biocompatibility characteristics of learner by the present invention, so that specific aim establishes study archives, during to be again started up study, learning robot can be according to the study archives coupling learning course of individual differences, to carry out differentiation study and training, targetedly student is helped to lift language ability.The present invention learning robot easily pronounced by recording learning person mistake word and sentence, and the fallibility in dialogue training talks with the content that the scholars such as sentence do not grasp and establishes corresponding study archives for sample, and constantly improve learns archives and is stored with the biocompatibility characteristics relevance of learner in learning process.
Description
Technical field
The present invention relates to electronic technology field, the method, apparatus and learning robot of study language are especially related to.
Background technology
Learning robot can help student to carry out phonetic study：It can engage in the dialogue with student and train, help student to carry
High spoken and language application ability.But since the crowd of language learning is numerous, horizontal level is also different, even horizontal
Everyone Grasping level to same knowledge point is also multifarious in the suitable crowd of level.It can not learn by learning robot
Middle progress differentiation study and training, effectively can not help student to lift language by targetedly recommending learned lesson
Learning ability.
Therefore, the prior art could be improved.
The content of the invention
The main object of the present invention is offer a kind of interactive learning methods, device and learning robot, it is intended to is solved existing
The technical problem of differentiation study and training cannot be carried out by learning robot.
The present invention proposes a kind of method for learning language, including：
The biocompatibility characteristics of identification learning person；
Judge whether the study archives associated with above-mentioned biocompatibility characteristics；
If in the presence of according to above-mentioned study archives coupling learning course.
Preferably, after above-mentioned the step of judging whether the study archives associated with above-mentioned biocompatibility characteristics, also
Including：
If it is determined that being not present, the selective listing of learned lesson is sent to learner；
Receive the selection instruction that the learner makes according to above-mentioned selective listing；
Learned lesson is matched to above-mentioned learner according to above-mentioned selection instruction.
Preferably, it is above-mentioned to be matched learned lesson to after the step of above-mentioned learner according to above-mentioned selection instruction, also wrap
Include：
The learning process information of recording learning person；
According to the above-mentioned study archives of above-mentioned learning process information creating；
Store above-mentioned study archives.
Preferably, above-mentioned storage described the step of learning archives, including：
Obtain the biocompatibility characteristics of learner；
Above-mentioned biocompatibility characteristics and above-mentioned study profile associated are stored.
Preferably, the step of biocompatibility characteristics of above-mentioned acquisition learner, including：
Collect the voice messaging of learner；
The vocal print feature of learner is extracted by specific mode according to above-mentioned voice messaging.
Preferably, above-mentioned specific mode includes：Gauss hybrid models or gauss hybrid models-universal background model.
Preferably, it is above-mentioned to be matched learned lesson to after the step of learner according to the selection instruction, including：
Whether the current above-mentioned learned lesson of analysis is adapted with the learning level of learner；
If it is not, then adjust above-mentioned learned lesson.
Preferably, it is above-mentioned according to the step of above-mentioned study archives coupling learning course before, including：
Send the module selective listing in above-mentioned learned lesson；
Receive the module selection instruction that learner makes according to above-mentioned module selective listing；
Module learned lesson in above-mentioned study archives is transferred according to above-mentioned module selection instruction.
Present invention also offers a kind of device for learning language, including：
Identification module, the biocompatibility characteristics for identification learning person；
Judgment module, for judging whether the study archives associated with above-mentioned biocompatibility characteristics；
First matching module, if for there are associated study archives, according to above-mentioned study archives coupling learning class
Journey.
Preferably, the device of above-mentioned study language, including：
First sending module, for if it is determined that there is no associated study archives, learned lesson to be sent to learner
Selective listing；
First receiving module, the selection instruction made for receiving learner according to above-mentioned selective listing；
Second matching module, for being matched learned lesson to above-mentioned learner according to above-mentioned selection instruction.
Preferably, the device of above-mentioned study language, including：
Logging modle, the learning process information for recording learning person；
Creation module, for according to the above-mentioned study archives of above-mentioned learning process information creating；
Memory module, for storing above-mentioned study archives.
Preferably, above-mentioned memory module, including：
Acquiring unit, for obtaining the biocompatibility characteristics of learner；
Storage unit, for above-mentioned biocompatibility characteristics and above-mentioned study profile associated to be stored.
Preferably, above-mentioned acquiring unit, including：
Subelement is collected, for collecting the voice messaging of learner；
Subelement is extracted, for extracting the vocal print feature of learner by specific mode according to above-mentioned voice messaging.
Preferably, above-mentioned specific mode includes：Gauss hybrid models or gauss hybrid models-universal background model.
Preferably, the device of above-mentioned study language, including：
Analysis module, for analyzing whether current above-mentioned learned lesson is adapted with the learning level of learner；
Module is adjusted, if the learning level for learned lesson and learner is incompatible, adjusts above-mentioned learned lesson.
Preferably, the device of above-mentioned study language, including：
Second sending module, for sending the module selective listing in above-mentioned learned lesson；
Second receiving module, the module selection instruction made for receiving learner according to above-mentioned module selective listing；
Module is transferred, for transferring the module learned lesson in above-mentioned study archives according to above-mentioned module selection instruction.
Present invention also offers a kind of learning robot, including memory, processor and it is at least one be stored in it is above-mentioned
In memory and the application program performed by above-mentioned processor is configured as, above application program is configurable for performing above-mentioned
Study language method.
Advantageous effects of the present invention：The present invention is identified different according to the difference of the biocompatibility characteristics of learner
People, so as to specific aim establish study archives, so as to be again started up study when, learning robot can be according to individual differences
Archives coupling learning course is practised, to carry out differentiation study and training, targetedly helps student to lift language ability.This hair
Bright learning robot easily pronounced by recording learning person mistake word and sentence, and dialogue training in fallibility dialogue
The content that the scholars such as sentence do not grasp establishes corresponding study archives for sample, and constantly improve learns archives in learning process
And it is stored in system with the biocompatibility characteristics relevance of learner.
Brief description of the drawings
The method flow schematic diagram of Fig. 1 one embodiment of the invention learning language；
The method flow schematic diagram of Fig. 2 another embodiment of the present invention learning language；
The method Optimizing Flow schematic diagram of Fig. 3 another embodiment of the present invention learning language；
The flow diagram of step S12 in Fig. 4 one embodiment of the invention；
The flow diagram of step S100 in Fig. 5 one embodiment of the invention；
The method re-optimization flow diagram of Fig. 6 another embodiment of the present invention learning language；
The method flow schematic diagram of Fig. 7 further embodiment of this invention learning language；
The apparatus structure schematic diagram of Fig. 8 one embodiment of the invention learning language；
The apparatus structure schematic diagram of Fig. 9 another embodiment of the present invention learning language；
The installation optimization structure diagram of Figure 10 another embodiment of the present invention learning language；
The structure diagram of memory module in Figure 11 one embodiment of the invention；
The structure diagram of acquiring unit in Figure 12 one embodiment of the invention；
The device re-optimization structure diagram of Figure 13 another embodiment of the present invention learning language；
The apparatus structure schematic diagram of Figure 14 further embodiment of this invention learning language.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Embodiment
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
Reference Fig. 1, the method for one embodiment of the invention learning language, including：
S1：The biocompatibility characteristics of identification learning person.
Biocompatibility characteristics in this step include gait feature, iris feature, skin characteristic, face as feature and vocal print feature
Deng passing through above-mentioned one or more biological characteristics and carry out individuations and distinguish identification and authentications；The preferred comparative learning person of the present embodiment
Vocal print feature.
S2：Judge whether the study archives associated with above-mentioned biocompatibility characteristics.
Study archives in this step include the history learning record data of learner, such as：Pronunciation training, recording learning
Person easily pronounce mistake word and sentence；Dialogue training, the dialogue sentence of recording learning fallibility.Study archives are gone through with study
Journey constantly improve and data message of change in learning process.
S3：If in the presence of according to above-mentioned study archives coupling learning course.
Study archives in this step are there are individual difference, and according to the study matched learned lesson of archives, there is also individual
Otherness, so that learning robot can be according to the study archives coupling learning course of individual differences, to carry out differentiation
Study and training, targetedly help student to lift language ability.Such as：In pronunciation training, according to study archives for easy
Wrong word and sentence carry out special practice, then arrange less practice for grasping relatively good word and sentence, have slapped
New word and sentence is arranged to be learnt after holding；In dialogue training, the dialogue easily to malfunction is directed to according to study archives and is carried out
Specialized training, is then suitably reviewed for grasping relatively good dialogue, and relatively good rear peace is all grasped in the dialogue learnt
New dialogue course is arranged to be learnt.
Reference Fig. 2, the method for another embodiment of the present invention learning language,
After step S2, including：
S4：If it is determined that being not present, the selective listing of learned lesson is sent to learner.
This step is by sending the selective listing of learned lesson, so as to controlling oneself in the selection course to be learnt for learner
Hold.
S5：Receive the selection instruction that above-mentioned learner makes according to above-mentioned selective listing.
Selection instruction includes：Study module and/or learning Content chapters and sections of selection etc..
S6：Learned lesson is matched to above-mentioned learner according to above-mentioned selection instruction.
The present embodiment selects learned lesson by learner oneself, closer to the true language proficiency of learner, to carry
High user experience.
Other embodiments of the invention can also be accounted for by the use accounting data of analytic learning course learning person, Select to use
Learned lesson more highest than data defines the language proficiency of most of learners, such as the use of certain chapters and sections modular learning course
Accounting data highest；The use accounting data highest of the learned lesson of a certain language proficiency grade for another example.Above-mentioned use accounts for
The ratio for all learners learnt by learning robot of learner's accounting for referring to use this learned lesson than data, according to
The language proficiency of most of learners carries out the learned lesson that coupling learning person learns first, with reduce matched learned lesson with
There is the very big probability of error in the true horizon of learner.
With reference to Fig. 3, the optimization of the method for the study language of another embodiment of the present invention, after step S6, further includes：
S10：The learning process information of recording learning person.
Learning process information in this step, such as, which partial knowledge point is grasped completely in learning process, which is partly known
Know the learning process information such as point is not grasped, which partial knowledge point study accumulated time is grown.Essential record learnt in the present embodiment
The knowledge point do not grasped in journey, such as：Recording learning person easily pronounce mistake word and sentence, the dialogue of recording learning fallibility
The content information of the fallibilities such as sentence.
S11：According to the above-mentioned study archives of above-mentioned learning process information creating.
By the acquisition study situation information in learning process information in this step, including to mastery of knowledge quality degree
Deng the above-mentioned study archives of establishment.The present embodiment learning person, which reviews process and new learning process, can add to the learner
Practise in archives, study archives can all change after study every time.
S12：Store above-mentioned study archives.
The study archives of this step may be selected to be stored in network system or learning robot service system, and the present embodiment is excellent
Choosing is stored in network system, to save the memory of learning robot, improves data processing speed.
Reference Fig. 4, the method for one embodiment of the invention learning language, step S12, including：
S100：Obtain the biocompatibility characteristics of learner.
The acquisition of this step refers to the biocompatibility characteristics identifying system capture by learning robot and recording learning person's correlation
Biocompatibility characteristics.It is preferred in the present embodiment to obtain the vocal print feature of learner, and corresponding characteristic parameter is obtained, such as：Sound
Short-time average energy, short-time magnitude and short-time energy frequency parameter of line etc.；The coefficient in Speech processing is additionally included, than
Such as：Derived characteristic parameter etc. under characteristic coefficient, auditory system model under sonification system model.The present embodiment passes through learning machine
The voice messaging of device people collection learner when being first learner's service, further to extract the vocal print feature of learner.
S101：Above-mentioned biocompatibility characteristics and above-mentioned study profile associated are stored.
The biocompatibility characteristics of learner and above-mentioned study archives are associated storage in this step, so as to learner again
It is recalled from network system according to biocompatibility characteristics during study and learn archives, and individual difference is further had according to learner
Study archives provide differentiation language learning service.
Reference Fig. 5, the method for one embodiment of the invention learning language, step S100, including：
S1001：Collect the voice messaging of learner.
The Application on Voiceprint Recognition of the present embodiment is divided into training stage and cognitive phase.In the training stage, phonetic study machine people system
System collects the training language section for some voice messagings that each learner says, and learning robot system carries out above-mentioned trained language section
Digitized processing, the template or model reference collection of each user are established according to characteristic parameter.
S1002：The vocal print feature of learner is extracted by specific mode according to above-mentioned voice messaging.
The present embodiment is trained voice using GMM-UBM (gauss hybrid models-universal background model) model, extraction
Vocal print feature, and preserve its corresponding characteristic parameter.Gauss hybrid models are selected to carry out voice in other embodiments of the invention
Training, extracts vocal print feature.
With reference to Fig. 6, the re-optimization of the method for another embodiment of the present invention learning language, after step S6, including：
S13：Whether the current above-mentioned learned lesson of analysis is adapted with the learning level of learner.
This step refer to learner for the first time using learning robot carry out phonetic study when, by being recorded during analytic learning
Study situation information analyze whether current above-mentioned learned lesson is adapted with the learning level of learner, such as：Word, sentence
The study situation information such as son and the accuracy of oral communication dialogue.Such as accuracy assert current above-mentioned between 70-90%
Practise whether course is adapted with the learning level of learner；Accuracy is less than 70% or more than 90%, then assert current above-mentioned
Whether incompatible with the learning level of learner practise course.
S14：If it is not, then adjust above-mentioned learned lesson.
This step refers to according to above-mentioned analysis situation, corresponding regularized learning algorithm course.Such as：Above-mentioned accuracy is less than 70%, then
It is adjusted to the corresponding low learned lesson of difficulty；Above-mentioned accuracy is more than 90%, then is adjusted to the corresponding high learned lesson of difficulty, with
Just the learned lesson recommended first improves the learning experience first of learner as close as the true horizon of learner.
With reference to Fig. 7, the method for further embodiment of this invention learning language, before step S3, including：
S30：Send the module selective listing in above-mentioned learned lesson.
There is Module Division function in the learning robot system of the present embodiment, such as：Word pronunciation learning module, dialogue study mould
Block etc., further to refine the study individuation difference of learner.Learner can lack according to the hobby or the study of itself of itself
Fall into by selecting respective modules in module selective listing, targetedly learnt.
S31：Receive the module selection instruction that learner makes according to above-mentioned module selective listing.
The selection instruction of this step includes clicking on by mouse or the selection signal of touch screen touch control manner triggering generation.
S32：Module learned lesson in above-mentioned study archives is transferred according to above-mentioned module selection instruction.
This step can form the differentiation study archives more refined by selecting module learned lesson in learning process.
Such as：In pronunciation training, learning robot reads word or sentence, and learner repeats robot and reads content, Learning machine
The pronunciation that people receives after the voice of learner with standard is contrasted, and prompts comparing result, while the pronunciation of recording learning person
Training is into study archives；In dialogue training, learning robot engages in the dialogue with learner and exchanges, and learning robot connects
Receive the voice of learner and judge whether learner answers correct by speech recognition, and provide prompting.Dialogue is learnt at the same time
Situation recorded in study archives.Recording learning archives are distinguished according to learner's difference vocal print feature, it is main to include easily pronunciation
The word and sentence of mistake, oral communication easily dialogue of mistake etc..Speech recognition in other embodiments of the invention uses
GMM-HMM methods or using deep approach of learning.
With reference to Fig. 8, the device of the study language of one embodiment of the invention, the device of above-mentioned study language is integrated in learning machine
On device people, including：
Identification module 1, the biocompatibility characteristics for identification learning person.
Biocompatibility characteristics in the present embodiment include gait feature, iris feature, skin characteristic, face as feature and vocal print are special
Sign etc., identifies that above-mentioned one or more biological characteristics carry out individuation and distinguish identification and authentication by identification module 1；The present embodiment
It is preferred that the vocal print feature by 1 comparative learning person of identification module.
Judgment module 2, for judging whether the study archives associated with above-mentioned biocompatibility characteristics.
Study archives in the present embodiment include the history learning record data of learner, such as：Pronunciation training, record are learned
Habit person easily pronounce mistake word and sentence；Dialogue training, the dialogue sentence of recording learning fallibility.Learn archives with study
Course constantly improve and data message of change in learning process.
First matching module 3, if for there are associated study archives, according to above-mentioned study archives coupling learning class
Journey.
Study archives in the present embodiment are there are individual difference, and the first matching module 3 is according to study matched of archives
Course is practised there is also individual difference, so that learning robot can be according to the study archives coupling learning class of individual differences
Journey, to carry out differentiation study and training, targetedly helps student to lift language ability.Such as：In pronunciation training, according to
Learn archives and carry out special practice for easily wrong word and sentence, for grasp relatively good word and sentence then arrange compared with
Few practice, arranges new word and sentence to be learnt after having grasped；In dialogue training, according to study archives for easy
The dialogue of error conducts special training, and is then suitably reviewed for grasping relatively good dialogue, the dialogue learnt is all
The new dialogue course of relatively good rear arrangement is grasped to be learnt.
With reference to Fig. 9, the device of the study language of another embodiment of the present invention includes：
First sending module 4, for if it is determined that there is no associated study archives, learned lesson to be sent to learner
Selective listing.
The present embodiment sends the selective listing of learned lesson by the first sending module 4, so as to the selection of controlling oneself of learner
The course content to be learnt.
First receiving module 5, the selection instruction made for receiving learner according to above-mentioned selective listing.
Selection instruction includes：Study module and/or learning Content chapters and sections of selection etc..
Second matching module 6, for being matched learned lesson to above-mentioned learner according to above-mentioned selection instruction.
The present embodiment carries out coupling learning course by the second matching module 6 according to the selection of learner oneself, so as to more
Close to the true language proficiency of learner, user experience is improved.
Other embodiments of the invention can also be accounted for by the use accounting data of analytic learning course learning person, Select to use
Learned lesson more highest than data defines the language proficiency of most of learners, such as the use of certain chapters and sections modular learning course
Accounting data highest；The use accounting data highest of the learned lesson of a certain language proficiency grade for another example.Above-mentioned use accounts for
The ratio for all learners learnt by learning robot of learner's accounting for referring to use this learned lesson than data, according to
The language proficiency of most of learners carries out the learned lesson that coupling learning person learns first, with reduce matched learned lesson with
There is the very big probability of error in the true horizon of learner.
Reference Figure 10, the optimization structure of the device of the study language of another embodiment of the present invention, including：
Logging modle 10, the learning process information for recording learning person.
By 10 recording learning procedural information of logging modle in the present embodiment, such as, which is partly known during recording learning
Know the learning process information such as point has been grasped completely, which partial knowledge point is not grasped, any partial knowledge point study accumulated time length.This
The knowledge point do not grasped in essential record learning process in embodiment, such as：Recording learning person easily pronounce mistake word and
The content information for talking with the fallibilities such as sentence of sentence, recording learning fallibility.
Creation module 11, for according to the above-mentioned study archives of above-mentioned learning process information creating.
By creation module 11 according to the acquisition study situation information in learning process information in the present embodiment, including to knowing
Grasp quality degree of knowledge etc. creates above-mentioned study archives.The present embodiment learning person reviews process and new learning process and can mend
It is charged in the study archives of the learner, study archives can all change after study every time.
Memory module 12, for storing above-mentioned study archives.
The study archives of the present embodiment can be stored in network system or learning robot service system by memory module 12
In, the present embodiment is preferably stored in network system, to save the memory of learning robot, improves data processing speed.
Reference Figure 11, one embodiment of the invention memory module 12, including：
Acquiring unit 100, for obtaining the biocompatibility characteristics of learner.
The acquisition of the present embodiment refers to the associated biomolecule feature by the capture of acquiring unit 100 and recording learning person.This reality
The vocal print feature of preferred acquisition learner in example is applied, and obtains corresponding characteristic parameter, such as：The short-time average energy of vocal print,
Short-time magnitude and short-time energy frequency parameter etc.；The coefficient in Speech processing is additionally included, such as：Under sonification system model
Characteristic coefficient, derived characteristic parameter etc. under auditory system model.The present embodiment is study first by learning robot
The voice messaging of learner is collected when person services, further to extract the vocal print feature of learner.
Storage unit 101, for above-mentioned biocompatibility characteristics and above-mentioned study profile associated to be stored.
The biocompatibility characteristics of learner are associated with above-mentioned study archives by storage unit 101 in the present embodiment and are deposited
Storage, recalls it from network system according to biocompatibility characteristics when learning again so as to learner and learns archives, and further basis
Learner has the service that the study archives of individual difference provide differentiation language learning.
Reference Figure 12, the acquiring unit 100 of one embodiment of the invention, including：
Subelement 1001 is collected, for collecting the voice messaging of learner.
The Application on Voiceprint Recognition of the present embodiment is divided into training stage and cognitive phase.In the training stage, by collecting subelement
1001 collect the training language section for some voice messagings that each learners say, learning robot system to above-mentioned trained language section into
Digitized processing, the template or model reference collection of each user are established according to characteristic parameter.
Subelement 1002 is extracted, for extracting the vocal print feature of learner by specific mode according to above-mentioned voice messaging.
Subelement 1002 is extracted in the present embodiment and uses GMM-UBM (gauss hybrid models-universal background model) model pair
Voice is trained, and extracts vocal print feature, and preserve its corresponding characteristic parameter.Gauss is selected to mix in other embodiments of the invention
Molding type is trained voice, extracts vocal print feature.
Reference Figure 13, the re-optimization structure of the device of another embodiment of the present invention learning language, including：
Analysis module 13, for analyzing whether current above-mentioned learned lesson is adapted with the learning level of learner.
The present embodiment passes through 13 analytics of analysis module when learner carries out phonetic study using learning robot for the first time
The study situation information recorded during habit analyzes whether current above-mentioned learned lesson is adapted with the learning level of learner,
Such as：The study situation information such as accuracy of word, sentence and oral communication dialogue.Such as accuracy is recognized between 70-90%
Whether settled preceding above-mentioned learned lesson is adapted with the learning level of learner；Accuracy is less than 70% or more than 90%, then recognizes
Whether settled preceding above-mentioned learned lesson is incompatible with the learning level of learner.
Module 14 is adjusted, if the learning level for learned lesson and learner is incompatible, adjusts above-mentioned study class
Journey.
The adjustment module 14 of the present embodiment is according to above-mentioned analysis situation, corresponding regularized learning algorithm course.Such as：Above-mentioned accuracy
Less than 70%, then the corresponding low learned lesson of difficulty is adjusted to；Above-mentioned accuracy is more than 90%, then it is corresponding high to be adjusted to difficulty
Learned lesson, the learned lesson to recommend first improve learner first as close as the true horizon of learner
Learning experience.
Reference Figure 14, the device of the study language of further embodiment of this invention, including：
Second sending module 30, for sending the module selective listing in above-mentioned learned lesson.
There is Module Division function in the language learning of the present embodiment, such as：Word pronunciation learning module, dialogue study module etc.,
Further to refine the study individuation difference of learner.The present embodiment is by the second sending module 30 to learner's sending module
Selective listing, so that learner can pass through the selection pair in module selective listing according to the hobby or the study defect of itself of itself
Module is answered, is targetedly learnt.
Second receiving module 31, the module selection instruction made for receiving learner according to above-mentioned module selective listing.
The selection instruction of the present embodiment includes the learner that the second receiving module 31 receives and passes through mouse click or touch screen
The selection signal of touch control manner triggering generation.
Module 32 is transferred, for transferring the module learned lesson in above-mentioned study archives according to above-mentioned module selection instruction.
The present embodiment transfers module learned lesson by transferring module 32, and the differentiation more refined is formed in learning process
Learn archives.Such as：In pronunciation training, learning robot reads word or sentence, and learner repeats robot and reads content,
The pronunciation that learning robot receives after the voice of learner with standard is contrasted, and prompts comparing result, while recording learning
The pronunciation training situation of person is into study archives；In dialogue training, learning robot engages in the dialogue with learner and exchanges, study
Robot receives the voice of learner and judges whether learner answers correct by speech recognition, and provides prompting.At the same time will
Dialogue study situation recorded in study archives.Recording learning archives are distinguished according to learner's difference vocal print feature, are mainly included
The easily word of pronunciation mistake and sentence, oral communication easily dialogue of mistake etc..Voice in other embodiments of the invention is known
Cai Yong not GMM-HMM methods or using deep approach of learning.
The embodiment of the present invention additionally provides a kind of learning robot, including memory, processor and at least one is stored
In above-mentioned memory and the application program performed by above-mentioned processor is configured as, above application program is configurable for holding
The method of the above-mentioned study language of row.
Different people is identified according to the difference of the biocompatibility characteristics of learner for the learning robot of the embodiment of the present invention,
So as to specific aim establish study archives, so as to be again started up study when, learning robot can be according to the study of individual differences
Archives coupling learning course, to carry out differentiation study and training, targetedly helps student to lift language ability.The present invention
The learning robot of embodiment easily pronounced by recording learning person mistake word and sentence, and dialogue training in fallibility
The content that the scholars such as dialogue sentence do not grasp establishes corresponding study archives for sample, and constantly improve learns in learning process
Archives are simultaneously stored in system with the biocompatibility characteristics relevance of learner.
The foregoing is merely the preferred embodiment of the present invention, is not intended to limit the scope of the invention, every utilization
The equivalent structure or equivalent flow shift that description of the invention and accompanying drawing content are made, it is related to be directly or indirectly used in other
Technical field, be included within the scope of the present invention.
Claims (10)
- A kind of 1. method for learning language, it is characterised in that including：The biocompatibility characteristics of identification learning person；Judge whether the study archives associated with the biocompatibility characteristics；If in the presence of according to the study archives coupling learning course.
- 2. the method for study language according to claim 1, it is characterised in that described to judge whether and the biology After the step of study archives that property feature is associated, further include：If it is determined that being not present, the selective listing of learned lesson is sent to learner；Receive the selection instruction that the learner makes according to the selective listing；Learned lesson is matched to the learner according to the selection instruction.
- 3. the method for study language according to claim 2, it is characterised in that described to be learnt according to the selection instruction Course is matched to after the step of learner, is further included：The learning process information of recording learning person；Learn archives according to the learning process information creating；Store the study archives.
- 4. the method for study language according to claim 3, it is characterised in that the step of the storage study archives Suddenly, including：Obtain the biocompatibility characteristics of learner；By the biocompatibility characteristics and the study profile associated storage.
- 5. the method for study language according to claim 4, it is characterised in that the biocompatibility characteristics for obtaining learner The step of, including：Collect the voice messaging of learner；The vocal print feature of learner is extracted by specific mode according to the voice messaging.
- A kind of 6. device for learning language, it is characterised in that including：Identification module, the biocompatibility characteristics for identification learning person；Judgment module, for judging whether the study archives associated with the biocompatibility characteristics；First matching module, if for there are associated study archives, according to the study archives coupling learning course.
- 7. the device of study language according to claim 6, it is characterised in that further include：First sending module, for if it is determined that there is no associated study archives, the selection of learned lesson to be sent to learner List；First receiving module, the selection instruction made for receiving the learner according to the selective listing；Second matching module, for being matched learned lesson to the learner according to the selection instruction.
- 8. the device of study language according to claim 7, it is characterised in that further include：Logging modle, the learning process information for recording learning person；Creation module, for learning archives according to the learning process information creating；Memory module, for storing the study archives.
- 9. the device of study language according to claim 8, it is characterised in that the memory module, including：Acquiring unit, for obtaining the biocompatibility characteristics of learner；Storage unit, for the biocompatibility characteristics and the study profile associated to be stored.
- 10. a kind of learning robot, including memory, processor and at least one it be stored in the memory and be configured For the application program performed by the processor, it is characterised in that the application program is configurable for perform claim requirement The method of study language in 1-5 described in any one.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711132285.4A CN107945619A (en) | 2017-11-15 | 2017-11-15 | Learn the method, apparatus and learning robot of language |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711132285.4A CN107945619A (en) | 2017-11-15 | 2017-11-15 | Learn the method, apparatus and learning robot of language |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107945619A true CN107945619A (en) | 2018-04-20 |
Family
ID=61932361
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711132285.4A Pending CN107945619A (en) | 2017-11-15 | 2017-11-15 | Learn the method, apparatus and learning robot of language |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107945619A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108960157A (en) * | 2018-07-09 | 2018-12-07 | 广东小天才科技有限公司 | Man-machine interaction method based on intelligent desk lamp and intelligent desk lamp |
| CN109360458A (en) * | 2018-10-25 | 2019-02-19 | 重庆鲁班机器人技术研究院有限公司 | Interest assistant teaching method, device and robot |
| CN110598208A (en) * | 2019-08-14 | 2019-12-20 | 清华大学深圳研究生院 | AI/ML enhanced pronunciation course design and personalized exercise planning method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1703709A (en) * | 2002-10-09 | 2005-11-30 | 李暎姬 | Internet studying system and the studying method |
| CN101241699A (en) * | 2008-03-14 | 2008-08-13 | 北京交通大学 | A speaker identification system for remote Chinese teaching |
| US8408913B2 (en) * | 2004-01-08 | 2013-04-02 | Linguaversal, SL | System, method, computer program and data set which are intended to facilitate language learning by means of sound identification |
| CN105118354A (en) * | 2015-09-14 | 2015-12-02 | 百度在线网络技术（北京）有限公司 | Data processing method for language learning and device thereof |
| CN106054620A (en) * | 2015-04-01 | 2016-10-26 | 全智慧科技股份有限公司 | Smart control apparatus and system |
| CN107239935A (en) * | 2017-06-06 | 2017-10-10 | 合肥创旗信息科技有限公司 | A kind of information orientated higher education platform based on e-learning |
-
2017
- 2017-11-15 CN CN201711132285.4A patent/CN107945619A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1703709A (en) * | 2002-10-09 | 2005-11-30 | 李暎姬 | Internet studying system and the studying method |
| US8408913B2 (en) * | 2004-01-08 | 2013-04-02 | Linguaversal, SL | System, method, computer program and data set which are intended to facilitate language learning by means of sound identification |
| CN101241699A (en) * | 2008-03-14 | 2008-08-13 | 北京交通大学 | A speaker identification system for remote Chinese teaching |
| CN106054620A (en) * | 2015-04-01 | 2016-10-26 | 全智慧科技股份有限公司 | Smart control apparatus and system |
| CN105118354A (en) * | 2015-09-14 | 2015-12-02 | 百度在线网络技术（北京）有限公司 | Data processing method for language learning and device thereof |
| CN107239935A (en) * | 2017-06-06 | 2017-10-10 | 合肥创旗信息科技有限公司 | A kind of information orientated higher education platform based on e-learning |
Non-Patent Citations (1)
| Title |
|---|
| 茶马古道之孙小美: "《说说"薄荷阅读"》", 19 August 2017, HTTP://BLOG.SINA.COM.CN/S/BLOG_5388AD510102YOSL.HTML * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108960157A (en) * | 2018-07-09 | 2018-12-07 | 广东小天才科技有限公司 | Man-machine interaction method based on intelligent desk lamp and intelligent desk lamp |
| CN109360458A (en) * | 2018-10-25 | 2019-02-19 | 重庆鲁班机器人技术研究院有限公司 | Interest assistant teaching method, device and robot |
| CN110598208A (en) * | 2019-08-14 | 2019-12-20 | 清华大学深圳研究生院 | AI/ML enhanced pronunciation course design and personalized exercise planning method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110728997B (en) | Multi-modal depression detection system based on context awareness | |
| CN110334626B (en) | Online learning system based on emotional state | |
| CN110782921B (en) | Voice evaluation method and device, storage medium and electronic device | |
| CN107918821A (en) | Teachers ' classroom teaching process analysis method and system based on artificial intelligence technology | |
| CN110544481B (en) | S-T classification method and device based on voiceprint recognition and equipment terminal | |
| CN108182937A (en) | Keyword recognition method, device, equipment and storage medium | |
| CN107609736A (en) | A kind of teaching diagnostic analysis system and method for integrated application artificial intelligence technology | |
| CN106777257B (en) | Intelligent dialogue model construction system and method based on dialect | |
| CN110299142B (en) | Voiceprint recognition method and device based on network convergence | |
| CN109117952B (en) | Robot emotion cognition method based on deep learning | |
| CN106297826A (en) | Speech emotional identification system and method | |
| CN109545218A (en) | Voice recognition method and system | |
| CN107945619A (en) | Learn the method, apparatus and learning robot of language | |
| CN113282840B (en) | Comprehensive training acquisition management platform | |
| CN110473548A (en) | A kind of classroom Internet analysis method based on acoustic signal | |
| CN110111223A (en) | Adaptive educational method and system based on artificial intelligence | |
| CN110797032A (en) | Voiceprint database establishing method and voiceprint identification method | |
| CN116259004B (en) | Student learning state detection method and system applied to online education | |
| CN116825288A (en) | Autism rehabilitation course recording method and device, electronic equipment and storage medium | |
| CN111405054A (en) | Intelligent teaching assistant system and method | |
| CN112052686B (en) | Voice learning resource pushing method for user interactive education | |
| CN114333787A (en) | Scoring method, device, equipment, storage medium and program product for spoken language examination | |
| CN108320592A (en) | A kind of autocontrol Oral English Practice pronunciation exercises system and method | |
| CN112863485A (en) | Accent voice recognition method, apparatus, device and storage medium | |
| CN114998968B (en) | Method for analyzing classroom interaction behavior based on audio and video |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180420 |