CN103458056A - Speech intention judging method based on automatic classification technology for automatic outbound system - Google Patents
Speech intention judging method based on automatic classification technology for automatic outbound system Download PDFInfo
- Publication number
- CN103458056A CN103458056A CN2013104377483A CN201310437748A CN103458056A CN 103458056 A CN103458056 A CN 103458056A CN 2013104377483 A CN2013104377483 A CN 2013104377483A CN 201310437748 A CN201310437748 A CN 201310437748A CN 103458056 A CN103458056 A CN 103458056A
- Authority
- CN
- China
- Prior art keywords
- intention
- automatic
- speech
- word
- user
- 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
- 238000000034 method Methods 0.000 title claims abstract description 54
- 238000005516 engineering process Methods 0.000 title claims abstract description 34
- 238000012549 training Methods 0.000 claims description 16
- 238000013507 mapping Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 9
- 238000006243 chemical reaction Methods 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 7
- 235000021167 banquet Nutrition 0.000 claims description 4
- 230000000694 effects Effects 0.000 claims description 3
- 230000008676 import Effects 0.000 claims description 3
- 238000004891 communication Methods 0.000 abstract 1
- 230000000875 corresponding effect Effects 0.000 description 7
- 238000003058 natural language processing Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 239000008358 core component Substances 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Abstract
The invention discloses a speech intention judging method based on an automatic classification technology for an automatic outbound system. The speech intention judging method is newly established by using a speech recognition technology and the automatic classification technology to judge the intention contained in the speeches of users scientifically and efficiently, the fact that classification of the intentions of users is completed based on an intermediate result - word and image network of the speech recognition is achieved, intention judging services are guaranteed to be carried out effectively in automatic outbound application scenes which are large in user speech accent, large in noise, high in narrow-band distortion rate and nonstandard in speech, the automatic outbound system is made to be capable of adopting proper response strategies to carry out conversations with the users according to the intentions of the users, and smooth communication between computers and the users is achieved eventually.
Description
Technical field
The present invention relates to the human-computer dialogue field, particularly relate to the voice intention decision method of a kind of automatic outer call system based on automatic classification technology.
Background technology
Outer paging system is the conventional business in call center, initiatively dials Subscriber Number by attending a banquet, with user session, to complete the particular tasks such as marketing, investigation.The automatic outer call system refers to that adopting computer to substitute attends a banquet, and with computer, dials Subscriber Number also and user session, completes a kind of application system of particular task.
In the automatic outer call system, computer will and complete particular task with user session, and one of them core technology is that the voice intention is judged, allows computer after receiving user speech, can judge by voice user's intention.
The judgement of voice intention refers to by computer processes processing for user's voice, the processing procedure of judgement user expressed intention in voice.It is the important step of automatic outer call system.High performance voice intention decision method, can guarantee to greatest extent the intention that interactive system can the correct understanding user, and take corresponding acknowledgment strategy, makes the dialog procedure between computer and user correctly and glibly to carry out.
In traditional voice intention judgement task, usually comprise two of speech recognitions, language analysis stage independently, by speech recognition technology, voice are converted into to word in the first stage, second stage is by natural language processing technique analysis user intention from word.
Speech recognition technology is the basic technology that the voice intention is judged, realizes the conversion from voice to the word.In recent years, the speech recognition technology development is very fast, and recognition capability is greatly improved.Along with the rise of mobile Internet, speech recognition technology has also progressed into application market, becomes mobile application another entrance except the word input.Yet these application all are confined to the application scenarios that the user coordinates, for stability and the input quality of user's phonetic entry, all have higher requirements.And in the field, call center, the characteristics such as lack of standardization because user speech exists that accent is large, noise is large, the arrowband distortion rate is high, the user speaks, the result of speech recognition rests on a very low level always, and this has limited the generation of automatic outer call system to a great extent.
Natural language processing is the key technology that the voice intention is judged.It can the metalanguage word, and in the identifiable language word, the user wishes the intent information of transmitting.Reaching the basis of this purpose, is to comprise correct user view in spoken and written languages.
In the field, call center, due to the difficulty reason of voice recognition tasks, make the word obtained from speech conversion, there is a large amount of mistakes, these error concealments user's intention, make follow-up natural language processing process, often can not get correct user view.
Automatic classification is a kind of most widely used mode identification method, and it mainly is divided into grader training and grader application two parts.In the grader training stage, can extract the mapping relations between input feature vector and output classification, build grader.In the grader application stage, can, according to the feature of new input, judge the target classification.Carry out the classifying space modeling because it adopts similarity to calculate, thereby there is stronger fault-tolerance, be suitable for the application scenarios of low quality for input feature vector.
The word that the error rate obtained for speech recognition is higher, it is return back to the word figure network that amount of information is larger, abandoning the method for traditional natural Language Processing analysis user intention simultaneously, then adopt the stronger sorting technique of fault-tolerance to carry out the user view judgement, is crucial innovative point of the present invention.
Summary of the invention
the actual techniques problem that the present invention solves: in order to solve that accuracy of speech recognition in the automatic outer call system is not high and user view that cause is judged failure problem, the invention provides a kind of voice intention decision method based on automatic classification technology, thereby guarantee better to judge by voice the accuracy of user view.
technical scheme of the present invention:the voice intention decision method of automatic outer call system based on automatic classification technology, it comprises
Cloud server, it is connected with the result database server with the automatic outer call system server, and cloud server contains task scheduling modules, sound identification module and intention sort module;
Task scheduling modules, its effect is the concurrent processing ability that promotes cloud server, by pressure, controls and load balancing, and cloud server can be realized the intention of multichannel user speech is judged simultaneously;
Sound identification module, receive the voice messaging that the automatic outer call system server imports into, the input voice are carried out to noise reduction, feature extraction processing, to input voice and be converted into characteristic sequence, then adopt the Hidden Markov continuous speech voice recognition device to carry out speech recognition, the phonetic feature sequence is converted into to the word figure network with confidence level;
The intention sort module, receive sound identification module and process rear information, word figure network is converted into to the space vector of Weight, and utilize the intent classifier trained in the intention sort module to be intended to automatic classification, the conversion of realization from voice to intention, finally according to intention institute corresponding character repertoire field, select the most correct word, form the speech recognition of high-accuracy, and the information afterwards of identifying is sent to the automatic outer call system server;
The result database server, for storing the cloud server operation result;
The automatic outer call system server, adopt computer to substitute and attend a banquet, and with computer, dials Subscriber Number also and user session, completes a kind of application system of particular task.
The voice intention of above-mentioned automatic outer call system based on automatic classification technology judged recognition methods, and the specific practice of training intent classifier comprises the following steps
A collects and intercepts the voice that comprise in a large number user view, manually marks the user view classification comprised in user speech, obtains in a large number the user speech with intentional mark;
B will, with the user speech of intentional mark, by speech recognition technology, be converted into the word figure network with confidence level by each user speech.
C, for word figure network, adopts feature selecting and feature extraction technology, selects most representative feature from word figure network, and realizes the mapping from word figure network to characteristic vector;
D is for the characteristic vector obtained from each user speech, itself and corresponding user view classification are carried out associated, and training mapping model between the two, i.e. intent classifier model.
The voice intention of aforesaid automatic outer call system based on automatic classification technology judged recognition methods, after often obtaining a user speech, the intention sort module can be to the further refinement of user view, provide information more accurately for automatic outer call system server and user link up, make to link up more natural, smooth.
The voice intention of aforesaid automatic outer call system based on automatic classification technology judged recognition methods, it is characterized in that: do not produce final recognition result with the speech recognition mould, but for obtaining the word figure network of Weight.
compared with the prior art, method provided by the invention has following beneficial effect:
(1) because voice identification result accuracy in the automatic outer call application is not high, can't be at these containing adopting natural language processing technique analysis user intention above wrong word, make the automatic outer call system be difficult to judge user view, finally cause the dialogue between system and user to reach target.The present technique invention has proposed the automatic classification technology higher based on serious forgiveness and has substituted the method that natural language processing technique carries out the judgement of voice intention.Because automatic classification technology is ignored the structural relation between word, realize that user speech arrives the mapping between intention but adopt similarity to calculate, be conducive to reduce the interference of the word of wrong identification to user view.
(2) current, the application of most of speech recognitions, after all being centered around and obtaining Word message by speech recognition, used Word message to carry out follow-up processing and process.This point is when the Word message accuracy is higher, and the user view of transmitting due to Word message is enough correct, makes it more applicable.When the Word message discrimination is low, a large amount of useful informations is abandoned by the speech recognition search algorithmic error in identifying, causes the user view of Word message transmission to be lost in identifying too.The word figure network that present technique invention has proposed to using in speech recognition process substitutes traditional Word message as recognition result, it can reduce the drop probability of speech recognition search algorithm to user intent information, makes user intent information be able to maximized preservation in the speech recognition stage.After word figure network result is delivered to the intention sort module, the user intent information that these preserve, will obtain correct judgement by intent classifier.
The accompanying drawing explanation
Fig. 1 is that the hardware topology schematic diagram of server disposition is beyond the clouds served in the intention judgement;
Fig. 2 is that intention is judged intention sort module training flow chart in service;
Fig. 3 is the word figure network configuration example that speech recognition produces;
Fig. 4 is that intention is judged the service applicating flow chart;
Fig. 5 is the execution mode architecture block diagram of the present invention in the automatic outer call system.
Embodiment
For understanding better the present invention, below in conjunction with embodiment and accompanying drawing, the present invention is more described in detail.In the following description, when perhaps the detailed description of existing prior art can desalinate subject content of the present invention, these are described in here and will be left in the basket.
accompanying drawing 1that the hardware topology schematic diagram of server disposition is beyond the clouds served in the intention judgement.
A kind of voice based on automatic classification technology provided by the invention are intended to decision method, can be used as the core component of automatic outer call system, judge the form of service with intention, are deployed on cloud server.
After deployment completes, cloud server has the ability of judging user view according to user speech: the automatic outer call system is submitted to cloud server by user speech, cloud server calls intention judgement service and carries out the user view judgement, and the user view that judgement is obtained returns to the automatic outer call system, correlated results is stored into to the result database system to back up simultaneously.
Server, comprise three modules: task scheduling modules, sound identification module and intention sort module beyond the clouds.Wherein,
Task scheduling modules, its effect is the concurrent processing ability that promotes cloud server, by pressure, controls and load balancing, and cloud server can be realized the intention of multichannel user speech is judged simultaneously;
Sound identification module, receive the voice messaging that the automatic outer call system server imports into, the input voice are carried out to noise reduction, feature extraction processing, to input voice and be converted into characteristic sequence, then adopt the Hidden Markov continuous speech voice recognition device to carry out speech recognition, the phonetic feature sequence is converted into to the word figure network with confidence level;
The intention sort module, receive sound identification module and process rear information, word figure network is converted into to the space vector of Weight, and utilize the intent classifier trained in the intention sort module to be intended to automatic classification, the conversion of realization from voice to intention, finally according to intention institute corresponding character repertoire field, select the most correct word, form the speech recognition of high-accuracy, and the information afterwards of identifying is sent to the automatic outer call system server;
accompanying drawing 2that intention is judged intention sort module training flow chart in service.
A kind of voice intention decision method based on automatic classification technology provided by the invention, broken through the method that traditional text according in user speech carrys out analysis user, but in the situation that speech recognition performance is poor, adopt not enough word figure network bearing results as speech recognition accurately, substitute with automatic classification method the method that text semantic is analyzed, realize the judgement to user view.
In the method, a basic job, be exactly the required grader of training intention sort module.The training step of grader is as follows:
(1) the training set is prepared in manual sorting, collects and intercept the voice that comprise in a large number user view, manually marks the user view classification comprised in user speech.Finally obtain judging the training set of service as intention with the user speech of intentional mark in a large number.
(2) to the user speech document in the training set, by speech recognition technology, each user speech is converted into to the word figure network with confidence level.
(3) for word figure network, adopt feature selecting and feature extraction technology, select most representative feature from word figure network, and realize the mapping from word figure network to characteristic vector.In the characteristic vector mapping process, can adopt various statistical values to weigh the amount of information of each word, comprise word document frequency (TF/IDF), information gain (Information Gain), mutual information (Mutual Information),
statistics (CHI) etc.Simultaneously, in these statistics, the word confidence level of word figure network is carried out to the statistical value stack in the mode of word frequency, make the characteristic vector of final realization can comprise the speech recognition confidence information.
(4) for the characteristic vector obtained from each user speech, itself and corresponding user view classification are carried out associated, and mapping model between the two of training, i.e. intent classifier model.At area of pattern recognition, existing a large amount of statistical classification (statistical classifiers) and machine learning (machine learning) method can be applied in the intention classified service, comprise: Support Vector Machine SVM(Support Vector Machines), k nearest neighbour classification kNN(k-Nearest Neighbor classifier), neural net method NNet(neural network), Naive Bayes Classification NB(Naive Bayes) method etc.
(5) final, utilize above method to complete training process, can obtain being applicable to the intent classifier model that intention is judged service.
accompanying drawing 3it is the word figure network configuration example that speech recognition produces.
The target of speech recognition is the conversion realized from the voice signal to the content of text.Because the speech recognition technology maturity is limit, in the call-center application scene, exist that accent is large, noise is large, the arrowband distortion rate is high, the user speaks characteristics such as lack of standardization, make the content of text that speech recognition finally obtains comprise more mistake.When these mistakes are carried out to text analyzing, can cause the misunderstanding of user view.
In fact, before speech recognition finally generates content of text, can, in its identifying, retain very abundant content information---word figure network.In word figure network, the nearly all possible content of text that voice are corresponding has all retained.Here at first can comprise more substantial wrong content, but also can comprise nearly all correct content simultaneously.These are wrong with together with correct content is obscured, and are difficult to practicality for text analyzing.
The present invention introduces the method for text classification innovatively, realizes the judgement to intention on word figure network.Because text classification is a kind of similarity calculating method in essence, do not need to resolve fully the fine structure in text, but all information fusions of obscuring can be calculated together, distinguish its weight according to various confidence level sum test statistics, the final judgement realized user view, thereby, under current speech recognition technology maturity, very strong adaptability is arranged.
The word figure network sample that is a user speech shown in figure.Can see,, in speech recognition word figure network, there is multiple possible recognition result in the words, and all correct contents are arranged, and has also comprised the content of a large amount of mistakes simultaneously.It all remains all information that should retain, and by follow-up intention classifying step, realizes the accurate assurance to intention.
accompanying drawing 4that intention is judged the service applicating flow chart.
A kind of voice intention decision method based on automatic classification technology provided by the invention, after the training stage completes, can realize the intention of arbitrary user speech is judged.Concrete steps are as follows:
(1) disposed intention and judged that the cloud server of service receives the voice that need to judge user view;
(2) user speech is submitted to sound identification module, be translated into the word figure network with confidence level;
(3) word figure network being submitted to intent classifier is classified.Intent classifier is by steps such as feature selecting, feature extractions, and after word figure network is converted into to characteristic vector, calling classification device model carries out kind judging;
(4) after obtaining user view, realize the conversion from voice to intention, finally according to intention institute corresponding character repertoire field, select the most correct word, form the speech recognition of high-accuracy, and will identify rear information and send to the automatic outer call system server;
accompanying drawing 5the execution mode architecture block diagram of the present invention in the automatic outer call system.
A kind of voice intention decision method based on automatic classification technology provided by the invention, be mainly used in the automatic outer call system.In application process, intention service module cloud server need to be docked with voice exchange (PBX) and the computer telephone integration server (CTI) of automatic outer call system, realization is transmitted timely to the voice that produce in the user answer process, and realizes the timely passback to the intention result of determination.
In the a-road-through words process of automatic outer call system, exist a plurality of dialogue bouts between system and user.At each dialogue bout, the voice intention that system is submitted to cloud server by PBX by voice delivery is judged service, the voice intention judges that service is for user speech judgement user view, after the users intention, this user view is returned to PBX and CTI, the automatic outer call system takes suitable acknowledgment strategy and user to engage in the dialogue according to user view, finally realizes that the smoothness between computer and user is linked up.
Claims (4)
1. the voice of an automatic outer call system based on automatic classification technology are intended to decision method, and it is characterized in that: it comprises
Cloud server, it is connected with the result database server with the automatic outer call system server, and cloud server contains task scheduling modules, sound identification module and intention sort module;
Task scheduling modules, its effect is the concurrent processing ability that promotes cloud server, by pressure, controls and load balancing, and cloud server can be realized the intention of multichannel user speech is judged simultaneously;
Sound identification module, receive the voice messaging that the automatic outer call system server imports into, the input voice are carried out to noise reduction, feature extraction processing, to input voice and be converted into characteristic sequence, then adopt the Hidden Markov continuous speech voice recognition device to carry out speech recognition, the phonetic feature sequence is converted into to the word figure network with confidence level;
The intention sort module, receive sound identification module and process rear information, word figure network is converted into to the space vector of Weight, and utilize the intent classifier trained in the intention sort module to be intended to automatic classification, the conversion of realization from voice to intention, finally according to intention institute corresponding character repertoire field, select the most correct word, form the speech recognition of high-accuracy, and the information afterwards of identifying is sent to the automatic outer call system server;
The result database server, for storing the cloud server operation result;
The automatic outer call system server, adopt computer to substitute and attend a banquet, and with computer, dials Subscriber Number also and user session, completes a kind of application system of particular task.
2. the voice intention of automatic outer call system according to claim 1 based on automatic classification technology judged recognition methods, and it is characterized in that: the specific practice of training intent classifier comprises the following steps
A collects and intercepts the voice that comprise in a large number user view, manually marks the user view classification comprised in user speech, obtains in a large number the user speech with intentional mark;
B will, with the user speech of intentional mark, by speech recognition technology, be converted into the word figure network with confidence level by each user speech;
C, for word figure network, adopts feature selecting and feature extraction technology, selects most representative feature from word figure network, and realizes the mapping from word figure network to characteristic vector;
D is for the characteristic vector obtained from each user speech, itself and corresponding user view classification are carried out associated, and training mapping model between the two, i.e. intent classifier model.
3. the voice intention of automatic outer call system according to claim 1 based on automatic classification technology judged recognition methods, it is characterized in that: after often obtaining a user speech, the intention sort module can be to the further refinement of user view, provide information more accurately for automatic outer call system server and user link up, make to link up more natural, smooth.
4. the voice intention of automatic outer call system according to claim 4 based on automatic classification technology judged recognition methods, it is characterized in that: do not produce final recognition result with the speech recognition mould, but for obtaining the word figure network of Weight.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310437748.3A CN103458056B (en) | 2013-09-24 | 2013-09-24 | Speech intention judging system based on automatic classification technology for automatic outbound system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310437748.3A CN103458056B (en) | 2013-09-24 | 2013-09-24 | Speech intention judging system based on automatic classification technology for automatic outbound system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103458056A true CN103458056A (en) | 2013-12-18 |
| CN103458056B CN103458056B (en) | 2017-04-26 |
Family
ID=49739998
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310437748.3A Active CN103458056B (en) | 2013-09-24 | 2013-09-24 | Speech intention judging system based on automatic classification technology for automatic outbound system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103458056B (en) |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104751843A (en) * | 2013-12-25 | 2015-07-01 | 上海博泰悦臻网络技术服务有限公司 | Voice service switching method and voice service switching system |
| CN105160229A (en) * | 2015-09-01 | 2015-12-16 | 武汉同迅智能科技有限公司 | Single-soldier system with voice and fingerprint dual authentication |
| CN105677779A (en) * | 2015-12-30 | 2016-06-15 | 山东大学 | Feedback-type question type classifier system based on scoring mechanism and working method thereof |
| CN106297771A (en) * | 2016-08-15 | 2017-01-04 | 烟台冰科医疗科技股份有限公司 | Multifunctional calling extension set |
| CN107924680A (en) * | 2015-08-17 | 2018-04-17 | 三菱电机株式会社 | Speech understanding system |
| CN108111491A (en) * | 2017-12-07 | 2018-06-01 | 浙江大学 | A kind of cell phone application voice storage management system based on artificial intelligence |
| CN108428447A (en) * | 2018-06-19 | 2018-08-21 | 科大讯飞股份有限公司 | A kind of speech intention recognition methods and device |
| CN108648752A (en) * | 2018-04-17 | 2018-10-12 | 重庆物奇科技有限公司 | A kind of intelligent sound control system and its control method based on cloud processing |
| CN109036431A (en) * | 2018-07-11 | 2018-12-18 | 北京智能管家科技有限公司 | A kind of speech recognition system and method |
| CN109104534A (en) * | 2018-10-22 | 2018-12-28 | 北京智合大方科技有限公司 | A kind of system for improving outgoing call robot and being intended to Detection accuracy, recall rate |
| CN109493850A (en) * | 2017-09-13 | 2019-03-19 | 株式会社日立制作所 | Growing Interface |
| CN109618067A (en) * | 2018-11-30 | 2019-04-12 | 携程旅游信息技术（上海）有限公司 | Outgoing call dialog process method and system |
| CN109893095A (en) * | 2019-03-11 | 2019-06-18 | 常州市贝叶斯智能科技有限公司 | A kind of intelligent robot system of body composition detection and analysis |
| CN109979453A (en) * | 2019-03-29 | 2019-07-05 | 客如云科技（成都）有限责任公司 | A kind of intelligent intention assessment man-machine interaction method towards the robot that orders |
| CN110233946A (en) * | 2019-06-17 | 2019-09-13 | 三角兽（北京）科技有限公司 | Execute outbound call service method, electronic equipment and computer readable storage medium |
| CN111290796A (en) * | 2018-12-07 | 2020-06-16 | 阿里巴巴集团控股有限公司 | Service providing method, device and equipment |
| CN111462726A (en) * | 2020-03-30 | 2020-07-28 | 中国建设银行股份有限公司 | Outbound response method, device, equipment and medium |
| CN111464707A (en) * | 2020-03-30 | 2020-07-28 | 中国建设银行股份有限公司 | Outbound call processing method, device and system |
| CN111475206A (en) * | 2019-01-04 | 2020-07-31 | 优奈柯恩(北京)科技有限公司 | Method and apparatus for waking up wearable device |
| CN111613034A (en) * | 2020-04-25 | 2020-09-01 | 国泰瑞安股份有限公司 | Fire-fighting monitoring control method and system |
| CN112331207A (en) * | 2020-09-30 | 2021-02-05 | 音数汇元(上海)智能科技有限公司 | Service content monitoring method and device, electronic equipment and storage medium |
| US20210110820A1 (en) * | 2019-10-10 | 2021-04-15 | Samsung Electronics Co., Ltd. | Method and apparatus with speech processing |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20010023397A1 (en) * | 1999-12-28 | 2001-09-20 | Kazuhiko Tajima | Conversation processing apparatus, method therefor, and recording medium therefor |
| CN101847405A (en) * | 2009-03-23 | 2010-09-29 | 索尼公司 | Speech recognition equipment and method, language model generation device and method and program |
| CN103295585A (en) * | 2012-02-24 | 2013-09-11 | 北京英立讯科技有限公司 | Processing system and method identifying whether telephone automatic dialing is responded by real human |
-
2013
- 2013-09-24 CN CN201310437748.3A patent/CN103458056B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20010023397A1 (en) * | 1999-12-28 | 2001-09-20 | Kazuhiko Tajima | Conversation processing apparatus, method therefor, and recording medium therefor |
| CN101847405A (en) * | 2009-03-23 | 2010-09-29 | 索尼公司 | Speech recognition equipment and method, language model generation device and method and program |
| CN103295585A (en) * | 2012-02-24 | 2013-09-11 | 北京英立讯科技有限公司 | Processing system and method identifying whether telephone automatic dialing is responded by real human |
Non-Patent Citations (3)
| Title |
|---|
| ZHIJIE YAN, F.K. SOONG, RENHUA WANG: "Word Graph based Feature Enhancement for Noisy Speech Recognition", 《IEEE INTERNATIONAL CONFERENCE ON ACOUSTIC, SPEECH AND SIGNAL PROCESSING, ICASSP2007》, vol. 4, 20 April 2007 (2007-04-20), pages 373 - 376 * |
| 张宝奇: "《中国优秀硕士学位论文全文数据库（电子期刊）信息科技辑》", 31 March 2012, article "基于切分的汉语连续语音识别技术研究" * |
| 朱琦，酆广增，肖海勇: "基于模式识别的语音分类方法", 《南京邮电学院学报（自然科学版）》, vol. 20, no. 4, 31 December 2000 (2000-12-31) * |
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104751843A (en) * | 2013-12-25 | 2015-07-01 | 上海博泰悦臻网络技术服务有限公司 | Voice service switching method and voice service switching system |
| CN107924680A (en) * | 2015-08-17 | 2018-04-17 | 三菱电机株式会社 | Speech understanding system |
| CN105160229A (en) * | 2015-09-01 | 2015-12-16 | 武汉同迅智能科技有限公司 | Single-soldier system with voice and fingerprint dual authentication |
| CN105677779B (en) * | 2015-12-30 | 2018-10-30 | 山东大学 | A kind of feedback-type problem types classifier system and its working method based on scoring |
| CN105677779A (en) * | 2015-12-30 | 2016-06-15 | 山东大学 | Feedback-type question type classifier system based on scoring mechanism and working method thereof |
| CN106297771A (en) * | 2016-08-15 | 2017-01-04 | 烟台冰科医疗科技股份有限公司 | Multifunctional calling extension set |
| CN109493850A (en) * | 2017-09-13 | 2019-03-19 | 株式会社日立制作所 | Growing Interface |
| CN108111491A (en) * | 2017-12-07 | 2018-06-01 | 浙江大学 | A kind of cell phone application voice storage management system based on artificial intelligence |
| CN108111491B (en) * | 2017-12-07 | 2020-06-19 | 浙江大学 | Management system is accomodate to cell-phone APP pronunciation based on artificial intelligence |
| CN108648752A (en) * | 2018-04-17 | 2018-10-12 | 重庆物奇科技有限公司 | A kind of intelligent sound control system and its control method based on cloud processing |
| CN108428447A (en) * | 2018-06-19 | 2018-08-21 | 科大讯飞股份有限公司 | A kind of speech intention recognition methods and device |
| CN108428447B (en) * | 2018-06-19 | 2021-02-02 | 科大讯飞股份有限公司 | Voice intention recognition method and device |
| CN109036431A (en) * | 2018-07-11 | 2018-12-18 | 北京智能管家科技有限公司 | A kind of speech recognition system and method |
| CN109104534A (en) * | 2018-10-22 | 2018-12-28 | 北京智合大方科技有限公司 | A kind of system for improving outgoing call robot and being intended to Detection accuracy, recall rate |
| CN109618067A (en) * | 2018-11-30 | 2019-04-12 | 携程旅游信息技术（上海）有限公司 | Outgoing call dialog process method and system |
| CN111290796A (en) * | 2018-12-07 | 2020-06-16 | 阿里巴巴集团控股有限公司 | Service providing method, device and equipment |
| CN111475206A (en) * | 2019-01-04 | 2020-07-31 | 优奈柯恩(北京)科技有限公司 | Method and apparatus for waking up wearable device |
| CN109893095A (en) * | 2019-03-11 | 2019-06-18 | 常州市贝叶斯智能科技有限公司 | A kind of intelligent robot system of body composition detection and analysis |
| CN109979453A (en) * | 2019-03-29 | 2019-07-05 | 客如云科技（成都）有限责任公司 | A kind of intelligent intention assessment man-machine interaction method towards the robot that orders |
| CN110233946A (en) * | 2019-06-17 | 2019-09-13 | 三角兽（北京）科技有限公司 | Execute outbound call service method, electronic equipment and computer readable storage medium |
| US20210110820A1 (en) * | 2019-10-10 | 2021-04-15 | Samsung Electronics Co., Ltd. | Method and apparatus with speech processing |
| US11508369B2 (en) * | 2019-10-10 | 2022-11-22 | Samsung Electronics Co., Ltd. | Method and apparatus with speech processing |
| US11830493B2 (en) | 2019-10-10 | 2023-11-28 | Samsung Electronics Co., Ltd. | Method and apparatus with speech processing |
| CN111464707A (en) * | 2020-03-30 | 2020-07-28 | 中国建设银行股份有限公司 | Outbound call processing method, device and system |
| CN111462726A (en) * | 2020-03-30 | 2020-07-28 | 中国建设银行股份有限公司 | Outbound response method, device, equipment and medium |
| CN111462726B (en) * | 2020-03-30 | 2023-08-22 | 中国建设银行股份有限公司 | Method, device, equipment and medium for answering out call |
| CN111613034A (en) * | 2020-04-25 | 2020-09-01 | 国泰瑞安股份有限公司 | Fire-fighting monitoring control method and system |
| CN112331207A (en) * | 2020-09-30 | 2021-02-05 | 音数汇元(上海)智能科技有限公司 | Service content monitoring method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103458056B (en) | 2017-04-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103458056A (en) | Speech intention judging method based on automatic classification technology for automatic outbound system | |
| CN104462600B (en) | Realize the method and device that incoming call reason is classified automatically | |
| US11361770B2 (en) | Detecting user identity in shared audio source contexts | |
| US10789943B1 (en) | Proxy for selective use of human and artificial intelligence in a natural language understanding system | |
| CN103474065A (en) | Method for determining and recognizing voice intentions based on automatic classification technology | |
| CN106406806A (en) | A control method and device for intelligent apparatuses | |
| CN111666380A (en) | Intelligent calling method, device, equipment and medium | |
| CN110381221B (en) | Call processing method, device, system, equipment and computer storage medium | |
| US11816609B2 (en) | Intelligent task completion detection at a computing device | |
| CN110265032A (en) | Conferencing data analysis and processing method, device, computer equipment and storage medium | |
| CN109614464B (en) | Method and device for identifying business problems | |
| TW202036323A (en) | Intelligent online customer service convergence core system which can provide a reply or query to an intention generated through voice recognition and natural language processing procedures | |
| CN110704618A (en) | Method and device for determining standard problem corresponding to dialogue data | |
| CN112818109A (en) | Intelligent reply method, medium, device and computing equipment for mail | |
| CN112364622A (en) | Dialog text analysis method, dialog text analysis device, electronic device and storage medium | |
| WO2015188454A1 (en) | Method and device for quickly accessing ivr menu | |
| CN107731230A (en) | A kind of court's trial writing-record system and method | |
| KR102030551B1 (en) | Instant messenger driving apparatus and operating method thereof | |
| CN113726942A (en) | Intelligent telephone answering method, system, medium and electronic terminal | |
| US20200153965A1 (en) | Caller deflection and response system and method | |
| CN114065720A (en) | Conference summary generation method and device, storage medium and electronic equipment | |
| TWI734085B (en) | Dialogue system using intention detection ensemble learning and method thereof | |
| CN110047473B (en) | Man-machine cooperative interaction method and system | |
| CN113111157B (en) | Question-answer processing method, device, computer equipment and storage medium | |
| CN112002306B (en) | Speech class recognition method and device, electronic equipment and readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Address after: The Zhongguancun Guiyang science and Technology Park core area and Guizhou street 550018 city of Guiyang province Guiyang national hi tech Industrial Development Zone No. 188 Applicant after: SHIJIHENGTONG TECHNOLOGY CO.,LTD. Address before: 550022 Guiyang Province, Guizhou national hi tech Zone, Jinyang intellectual Industrial Park Venture Building, room 526 Applicant before: GUIYANG SHIJIHENGTONG TECHNOLOGY Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Li Jianzhou Inventor after: Yang Xingrong Inventor before: Li Jianzhou Inventor before: Yang Xingrong |
|
| CB03 | Change of inventor or designer information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20190215 Address after: Room 3F, Torch Software Park, Xintian Avenue Pioneering Road, Wudang District, Guiyang City, Guizhou Province Patentee after: Guiyang Fengchi Technology Co.,Ltd. Address before: 550018 No. 188 Nanyuan Science and Education Street, the core area of Guiyang Science and Technology Park, Zhongguancun, Guiyang National High-tech Industrial Development Zone, Guiyang City, Guizhou Province Patentee before: SHIJIHENGTONG TECHNOLOGY CO.,LTD. |
|
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: Speech intention judging system based on automatic classification technology for automatic outbound system Effective date of registration: 20190619 Granted publication date: 20170426 Pledgee: Bank of Guiyang Limited by Share Ltd. high tech branch Pledgor: Guiyang Fengchi Technology Co.,Ltd. Registration number: 2019520000010 |
|
| PE01 | Entry into force of the registration of the contract for pledge of patent right | ||
| PC01 | Cancellation of the registration of the contract for pledge of patent right |
Date of cancellation: 20190703 Granted publication date: 20170426 Pledgee: Bank of Guiyang Limited by Share Ltd. high tech branch Pledgor: Guiyang Fengchi Technology Co.,Ltd. Registration number: 2019520000010 |
|
| PC01 | Cancellation of the registration of the contract for pledge of patent right | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: Speech intention judging system based on automatic classification technology for automatic outbound system Effective date of registration: 20190715 Granted publication date: 20170426 Pledgee: Bank of Guiyang Limited by Share Ltd. high tech branch Pledgor: Guiyang Fengchi Technology Co.,Ltd. Registration number: 2019520000012 |
|
| PE01 | Entry into force of the registration of the contract for pledge of patent right | ||
| PC01 | Cancellation of the registration of the contract for pledge of patent right |
Date of cancellation: 20210607 Granted publication date: 20170426 Pledgee: Bank of Guiyang Limited by Share Ltd. high tech branch Pledgor: Guiyang Fengchi Technology Co.,Ltd. Registration number: 2019520000012 |
|
| PC01 | Cancellation of the registration of the contract for pledge of patent right | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: Automatic outbound call system based on automatic classification technology for speech intention determination system Effective date of registration: 20230713 Granted publication date: 20170426 Pledgee: China Construction Bank Corporation Guiyang Chengbei sub branch Pledgor: Guiyang Fengchi Technology Co.,Ltd. Registration number: Y2023980048209 |
|
| PE01 | Entry into force of the registration of the contract for pledge of patent right |