CN115938347B - Flight student conversation normalization scoring method and system based on voice recognition - Google Patents
Flight student conversation normalization scoring method and system based on voice recognition Download PDFInfo
- Publication number
- CN115938347B CN115938347B CN202310236989.5A CN202310236989A CN115938347B CN 115938347 B CN115938347 B CN 115938347B CN 202310236989 A CN202310236989 A CN 202310236989A CN 115938347 B CN115938347 B CN 115938347B
- Authority
- CN
- China
- Prior art keywords
- flight
- call
- text
- deviation
- student
- 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.)
- Active
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method and a system for scoring call standardability of flight students based on voice recognition. According to the method and the system, voice call content is obtained in real time and converted into text, calculation of call content deviation is completed based on semantic matching and comparison verification models, and finally, nonstandard labels and scores of calls of flight students in the flight training period are given, so that the flight students can recognize and correct the call problems, the cultivation degree of the flight students is improved, and the aviation safety problem caused by the land-air call quality can be reduced.
Description
Technical Field
The invention relates to the field of ground-air call normalization scoring, in particular to a method and a system for scoring flight student call normalization based on voice recognition.
Background
The training of air traffic control instruction standard expression is very important and indispensable content in air traffic management, and the training of land-air communication standard expression has historically been an important point in air controller culture. And the aircraft is operated safely and efficiently, and the air traffic controller and the pilot can accurately communicate and understand each other. Therefore, whether the land-air communication term of the pilot is standard and standard is also important to the influence of the flight safety.
With the rapid development of civil aviation technology, the faults of communication equipment are gradually reduced, but unsafe aviation events such as inaccurate repeated reading, misunderstanding, irregular words, incomplete content and the like related to wireless telephone communication errors still occur. More than 50% of aviation accidents are related to wireless telephone communication errors, as reported by NASA. With the continuous development of the civil aviation transportation industry in China, the air traffic flow is rapidly increased, the working pressure of pilots is also increased, the demands on high quality pilots are also increased, and higher demands are also put on the cultivation of flight students. During the flight of the flight students, especially during the flight learning period, tension, inexperience, huge pressure and the like, the energy is mainly concentrated on the operation, the rapid land-air communication instruction of the controllers is difficult to comprehensively cope with, and the phenomena of readback errors, irregular expression and the like are unavoidable.
Therefore, a specific set of problems capable of feeding back the conversation term during the flight learning of the flight student are required to be designed, the establishment of the standardability concept of the land-air conversation term is helped, the land-air conversation is standardized, and the flight safety of the flight student in the actual work in future is ensured.
Disclosure of Invention
The invention aims to provide a method and a system for scoring the standardability of the communication of an aeronaut based on voice recognition, which are used for standardizing the contents of the communication of the aeronaut and the air in the process of learning the aeronaut and reducing the problem of unsafe aviation caused by related language errors.
In order to achieve the above object, the present invention provides the following technical solutions:
a method for scoring call normalization of an aircraft student based on speech recognition, the scoring method comprising:
s1: acquiring the voice uploaded by the land-air communication terminal in real time;
s2: converting the speech to text;
s3: analyzing and classifying the text, and marking and positioning the communication text of the flight student;
s4: calculating the nonstandard deviation of the call text of the flight student based on the semantic matching and comparison verification model, assigning a heat value vector hvv, and giving a total value hvvM of the heat value vector on each flight ending the flight training;
s5: and calculating and evaluating the flight student conversation text according to the scoring key points and the total value hvvM of the heat value vector.
Preferably, S3 specifically includes:
s31: identifying the flight corresponding to the instruction from the call sign in the dialogue text;
s32: the instructions of the same call sign are arranged according to the time sequence;
s33: identifying a land-air call role through part-of-speech and semantic analysis;
s34: the flight student talk text is marked and located.
Preferably, S4 specifically includes:
s41: after the call text of the flight student is obtained, the content of the call text is taken as a training set, whether deviation exists in each section of call text of the flight student is marked by adopting a manual marking mode, and a heat value vector hvv is assigned according to a deviation result;
s42: putting the marked sample into the training of the semantic matching and contrast verification model;
s43: putting the trained semantic matching and comparison verification model into use, and directly outputting a deviation result and a heat value vector hvv of the communication text of the flight student by an algorithm;
s44: for each flight ending the flight training, the overall dimension of the hvvM matrix is n x, n is the number of rows, that is, the total number of calls of the flight student after the flight is ended, and x is the number of non-standard classifications for inclusion of the deviation score.
Preferably, the deviation in S4 includes a repeated inaccuracy deviation, a word non-standardization deviation, a content incompleteness deviation, and a misunderstanding deviation.
Preferably, S5 is calculated and evaluated according to the following formula:
e=a*Sum[hvvM[[:,x]]]/n*100%
s=100%-e
wherein e is the percentage of the deduction, a is the deviation definition weighting coefficient in the range of 0-2, s is the total score, x is the number of non-standard classifications for the inclusion deviation score, and n is the number of lines.
In order to achieve the above object, the present invention provides the following technical solutions:
a speech recognition-based flight attendant call normalization scoring system, comprising:
a voice unit: the method is used for acquiring the voice uploaded by the land-air communication terminal in real time;
conversion unit: for converting the speech to text;
analysis unit: the text analyzing and classifying device is used for analyzing and classifying the text and marking and positioning the call text of the flight student;
comparison unit: the method comprises the steps of calculating nonstandard deviation of call text of the flight student based on a semantic matching and comparison verification model, assigning a heat value vector hvv, and giving a total value hvvM of the heat value vector on each flight ending flight training;
an evaluation unit: and the method is used for calculating and evaluating the flight student conversation text according to the scoring key points and the total value hvvM of the heat value vector.
Preferably, the analysis unit specifically comprises:
and an identification module: the method comprises the steps of identifying an instruction corresponding flight from a call sign in a dialogue text;
and a sequencing module: the instruction is used for arranging the same call sign according to the time sequence;
and an analysis module: the method is used for identifying the ground-air call roles through part-of-speech and semantic analysis;
and a marking module: the method is used for marking and positioning the communication text of the flight student;
preferably, the comparing unit specifically includes:
and the marking module is used for: after the call text of the flight student is obtained, marking whether each section of call text of the flight student has deviation or not by taking the content of the call text as a training set in a manual marking mode, and assigning a heat value vector hvv according to a deviation result;
training module: the method is used for putting the marked sample into the training of the semantic matching and contrast verification model;
and an output module: the algorithm is used for putting the trained semantic matching and comparison verification model into use and directly outputting deviation results and heat value vectors hvv of the communication text of the flight students;
and a summarizing module: for each flight ending the flight training, the total dimension of the hvvM matrix is n x, n is the number of rows, that is, the total number of calls of the learner after the flight is ended, and x is the number of non-standard classifications for inclusion of the deviation score.
Preferably, the evaluation unit is adapted to calculate and evaluate according to the following formula:
e=a*Sum[hvvM[[:,x]]]/n*100%
s=100%-e
wherein e is the percentage of the deduction, a is the deviation definition weighting coefficient in the range of 0-2, s is the total score, x is the number of non-standard classifications for the inclusion deviation score, and n is the number of lines.
Preferably, the scoring system is able to recover call content and annotate unnormalized and correct phrase content in the scoring report and print the scoring report.
Compared with the prior art, the invention has the beneficial effects that:
the system for scoring the conversation standardability of the flight student based on the voice recognition can enable the flight student to realize the problem and correct the related nonstandard content through the scoring report and the conversation content restored and marked in the scoring report after finishing the flight training task of the current day, improves the cultivation degree of the flight student, and reduces the aviation safety problem caused by the land-air conversation quality; the scoring system starts to be started when the flight student executes single training flight, and an evaluation result is obtained after the single training flight is finished, so that the whole conversation quality evaluation process does not need manual intervention.
Drawings
FIG. 1 is a schematic diagram of a system operating environment of the present invention;
FIG. 2 is a flowchart of a scoring system of the present invention;
FIG. 3 is a diagram illustrating a call normalization bias identification and classification method according to the present invention;
fig. 4 is a call normalization scoring gist of the present invention.
Detailed Description
The present invention will be described in further detail with reference to test examples and specific embodiments. It should not be construed that the scope of the above subject matter of the present invention is limited to the following embodiments, and all techniques realized based on the present invention are within the scope of the present invention.
Example 1
According to the invention, through real-time voice recognition and semantic matching comparison verification of communication of flight students according to voice texts, comparison is carried out with standard land-air communication, and deviation values between various comparison factors of real-time voice and comparison factors of voice samples are obtained; finally, carrying out overall evaluation on real-time voice of the flight student according to the deviation value to obtain an evaluation total score, recovering all land-air calls of the training flight corresponding to the same call sign in a report form after the flight is finished, marking the deviation type of the part with deviation in the call of the flight student, and remarking the call content with correct standard. Therefore, the communication content of the flight student can be comprehensively monitored, reminded and corrected, so that the aim of improving the standardability of the land-air communication of the flight student is fulfilled.
Referring to fig. 1, which is a schematic diagram of an operation environment of a speech recognition-based flight student call normalization scoring system according to the present invention, the speech recognition-based flight student call normalization scoring system receives land-air call speech uploaded in real time from a flight student 1 and a flight director 2, and from a flight student 1 and a controller 3, wherein the flow direction of the speech is a report of a flow scoring result from the flight student 1 and the flight director 2, or from the flight student 1 and the controller 3 through a speech recognition-based flight student call normalization scoring system 4. The speech recognition and flight student conversation normalization scoring system 4 is a system core, performs real-time text conversion on speech in the conversation process, performs serial semantic matching comparison verification, and prints and outputs an analysis result to the terminal 5 so as to facilitate a pilot to timely normalize own land-air conversation after completing on-duty training flight.
Referring to fig. 2, a workflow diagram of each unit of the system for normalized scoring of a communication of an aeronaut based on speech recognition according to the present invention includes:
s1: based on the voice unit, the voices uploaded by the air-ground communication terminals 2 and 3 are acquired. The terminals 2 and 3 are communicated with the system in real time, and the communication voice can be automatically uploaded to the system as long as the land-air communication is carried out.
S2: based on the recognition unit, the speech is converted into text. The system converts speech to text in real time based on streaming speech recognition.
S3: based on the part-of-speech analysis of the control text of the analysis unit, different call signs correspond to instruction clusters and instruction roles are identified. According to text content, part-of-speech analysis is carried out, firstly, flights corresponding to instructions are identified according to call signs in dialogue texts, and the instructions of the same call sign are arranged and aggregated into a class according to time sequence; secondly, through part of speech and semantic analysis, the role of an issuer of the voice call is identified, which of an aeronaut, a flight director and a controller is determined, and the label is added, so that the aeronaut-air call content of the aeronaut is conveniently positioned.
S4: based on the semantic matching and comparison verification model of the comparison unit, training the semantic matching and comparison verification model, and importing and calculating the call deviation of each flight student. And (3) positioning call contents of flight students according to instruction clustering and instruction role identification marking results corresponding to different call signs in the step (S3), and identifying and classifying call deviations of the flight students by combining instruction contents of flight instructors and controllers in the immediately preceding and following time.
In one case of this embodiment, referring to fig. 3, a method for calculating a semantic matching and contrast verification model according to the present invention includes:
and (3) taking the call text of the ith flight student of each identified role and all call texts of the corresponding flights before the call text as input to a land-air call clustering result corresponding to each call sign, importing a semantic matching and contrast verification model, initializing a heat value vector (hvv), wherein the length of the heat value vector is 4 (the heat value vector is classified into 4 types of unnormals and can be expanded according to actual conditions), and the initial value is 0, namely hvv = {0, 0}. The specific method for identifying various deviations comprises the following steps:
s41: repeating the inaccurate deviation. Based on the semantic matching and contrast verification model, if the recognition result is the repeated reading, judging whether the repeated reading is accurate, if not, marking the word with repeated reading error, assigning the first element of hvv as 1, namely hvv = {1, 0}, ending the recognition of the communication text of the ith flight student, and integrating hvv = {1, 0} as the recognition result into a total pilot communication normalized heat value vector matrix (hot value vector Matrix: hvM), namely hvvM [ i ] = hvv.
S42: the term is used to denominate deviations. Based on the semantic matching and comparison verification model, if the recognition result is irregular, marking out irregular words, assigning a second element of hvv as 1, namely hvv = {0,1, 0}, ending the recognition of the ith flight student conversation text, and integrating hvv = {0,1, 0} as the recognition result into a total pilot conversation normalized heat value vector matrix (hot value vector Matrix: hvM), namely hvvM [ i ] = hvv.
S43: incomplete content deviations. Based on the semantic matching and comparison verification model, if the identification result is incomplete, marking words with incomplete contents, assigning a third element of hvv to be 1, namely hvv = {0,1, 0}, ending the identification of the ith flight student conversation text, and integrating hvv = {0,1, 0} as the identification result into a total pilot conversation normative heat value vector matrix (hot value vector Matrix: hvvM), namely hvvM [ i ] = hvv.
S44: the bias is misinterpreted. Based on the semantic matching and comparison verification model, if the recognition result is misunderstanding, misunderstanding words are marked, a fourth element of hvv is assigned to be 1, namely hvv = {0, 1}, the recognition of the ith flight student conversation text is ended, hvv = {0, 1} is taken as the recognition result, and the recognition result is integrated into a total pilot conversation normative heat value vector matrix (hot value vector Matrix: hvvM), namely hvvM [ i ] = hvv.
If the recognition result does not belong to the four types of deviation, the land-air call is indicated as standard, and the initial value, namely hvv = {0, 0}, is maintained.
For each flight ending the flight training, the overall dimension of the hvvM matrix is obtained as n×4, n is the number of rows, i.e. the total number of student calls after the end of the flight, and 4 is the 4 non-canonical classifications in this example.
S5: real-time speech is evaluated. And defining a weight coefficient for each deviation, calculating the percentage of the deduction of each deviation according to the total deviation vector obtained in the step S4, and obtaining the total deduction through summation, thereby obtaining the total score, wherein the specific method comprises the following steps:
referring to fig. 4, the call normalization scoring points of the flight student according to the embodiment of the invention include:
s51: the repeated reading is inaccurate, the percentage of the deduction e1 is as follows:
e1=a1*Sum[hvvM[[:,1]]]/n*100%
s52: the word is not standard, the percentage e2 is given by the formula:
e2=a2*Sum[hvvM[[:,2]]]/n*100%
s53: incomplete content, percentage of deduction e3, the formula is:
e3=a3*Sum[hvvM[[:,3]]]/n*100%
s54: misunderstanding, the percentage of the deduction e4 is as follows:
e4=a4*Sum[hvvM[[:,4]]]/n*100%
s55: the formula is:
s=100- e1- e2- e3- e4
wherein s is the total score, a 1-a 4 define a weighting coefficient for each deviation in the range of 0-2,
s6: the report with the total score and each deviation deduction is printed, the report contains the original call content, and places, deviation categories and correct term content of the communication of the flight student are marked.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, and alternatives falling within the spirit and principles of the invention.
Claims (7)
1. A method for scoring call normalization of an aircraft student based on voice recognition, the scoring method comprising:
s1: acquiring voice of service call;
s2: converting the speech to text;
s3: analyzing and classifying the text, and marking and positioning the communication text of the flight student therein;
s4: inputting the communication text of the flight student based on a trained semantic matching and comparison verification model, outputting nonstandard deviation of the communication text of the flight student, assigning a heat value vector hvv, integrating the heat value vector hvv into a heat value vector matrix hvM, and obtaining a heat value vector matrix hvM with the total dimension of n x on each flight ending the flight training, wherein n is the number of lines, and x is the nonstandard classification number;
s5: calculating and evaluating call texts of flight students according to the scoring key points and the heat value vector matrix hvvM, wherein when x=4, the nonstandard deviation comprises repeated reading inaccuracy deviation, word nonstandard deviation, incomplete content deviation and misunderstanding deviation, and the calculation and evaluation are performed according to the following formula:
s51: the repeated reading is inaccurate, the percentage of the deduction e1 is as follows:
e1=a1*Sum[hvvM[[:,1]]]/n*100%,
s52: the word is not standard, the percentage e2 is given by the formula:
e2=a2*Sum[hvvM[[:,2]]]/n*100%,
s53: incomplete content, percentage of deduction e3, the formula is:
e3=a3*Sum[hvvM[[:,3]]]/n*100%,
s54: misunderstanding, the percentage of the deduction e4 is as follows:
e4=a4*Sum[hvvM[[:,4]]]/n*100%,
s=100-e1-e2-e3-e4,
wherein a 1-a 4 define a weighting coefficient for deviations ranging from 0-2, s is the total score, and n is the number of rows.
2. The method for scoring call normalization of flight students based on voice recognition according to claim 1, wherein S3 specifically comprises:
s31: identifying the flight corresponding to the instruction from the call sign in the dialogue text;
s32: the instructions of the same call sign are arranged according to the time sequence;
s33: identifying a land-air call role through part-of-speech and semantic analysis;
s34: the flight student talk text is marked and located.
3. The method for scoring call normalization of flight students based on voice recognition according to claim 1, wherein S4 specifically comprises:
s41: after the call text of the flight student is obtained, the content of the call text is taken as a training set, whether each section of call text of the flight student has an irregular deviation or not is marked in a manual marking mode, and a heat value vector hvv is assigned according to a deviation result;
s42: putting the marked sample into the training of the semantic matching and contrast verification model;
s43: putting the trained semantic matching and comparison verification model into use, and directly outputting a deviation result and a heat value vector hvv of the communication text of the flight student by an algorithm;
s44: for each flight ending the flight training, the total dimension of the heat value vector matrix hvvM is n x, n is the number of rows, namely the total number of calls of the flight students after the flight is ended, and x is the number of non-standard classifications for inclusion of the deviation scoring item.
4. A speech recognition-based flight attendant call normalization scoring system, comprising:
a voice unit: the method is used for acquiring the voice uploaded by the service call terminal in real time;
conversion unit: for converting the speech to text;
analysis unit: the text is used for analyzing and classifying the text, and marking and positioning the communication text of the flight student therein;
comparison unit: the method comprises the steps of inputting a call text of a flight student based on a trained semantic matching and comparison verification model, outputting nonstandard deviation of the call text of the flight student, assigning a heat value vector hvv, integrating the heat value vector hvv into a heat value vector matrix hvvM, and obtaining a heat value vector matrix hvvM with a total dimension of n x from each flight ending flight training, wherein n is the number of rows, and x is the nonstandard classification number;
an evaluation unit: the method is used for calculating and evaluating the call text of the flight student according to the scoring key points and the heat value vector matrix hvvM, when x=4, the nonstandard deviation comprises repeated reading inaccuracy deviation, word nonstandard deviation, incomplete content deviation and misunderstanding deviation, and the calculation and evaluation are carried out according to the following formula:
s51: the repeated reading is inaccurate, the percentage of the deduction e1 is as follows:
e1=a1*Sum[hvvM[[:,1]]]/n*100%,
s52: the word is not standard, the percentage e2 is given by the formula:
e2=a2*Sum[hvvM[[:,2]]]/n*100%,
s53: incomplete content, percentage of deduction e3, the formula is:
e3=a3*Sum[hvvM[[:,3]]]/n*100%,
s54: misunderstanding, the percentage of the deduction e4 is as follows:
e4=a4*Sum[hvvM[[:,4]]]/n*100%,
s=100-e 1-e2-e3-e4, wherein a 1-a 4 define a weighting coefficient for deviations ranging from 0-2, s is the total score, and n is the number of lines.
5. The speech recognition-based flight attendant conversation normalization scoring system of claim 4, wherein said analysis unit comprises:
and an identification module: the method comprises the steps of identifying an instruction corresponding flight from a call sign in a dialogue text;
and a sequencing module: the instruction is used for arranging the same call sign according to the time sequence;
and an analysis module: the method is used for identifying the ground-air call roles through part-of-speech and semantic analysis;
and a marking module: for marking and locating the flight attendant call text therein.
6. The speech recognition-based flight attendant conversation normalization scoring system of claim 4 wherein said comparison means comprises:
and the marking module is used for: after the call text of the flight student is obtained, marking whether each section of call text of the flight student has an irregular deviation by taking the content of the call text as a training set in a manual marking mode, and assigning a heat value vector hvv according to a deviation result;
training module: the method is used for putting the marked sample into the training of the semantic matching and contrast verification model;
and an output module: the algorithm is used for putting the trained semantic matching and comparison verification model into use and directly outputting deviation results and heat value vectors hvv of the communication text of the flight students;
and a summarizing module: for each flight ending the flight training, the total dimension of the calorific value vector matrix hvvM is n x, n is the number of rows, that is, the total number of calls of the learner after the flight is ended, and x is the number of non-standard classifications for inclusion of the deviation score.
7. The speech recognition based flight student conversation normalization scoring system of claim 4 wherein the scoring system generates scoring reports after the assessment is complete, recovers conversation content and labels non-normative and correct term content, and prints the scoring reports.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310236989.5A CN115938347B (en) | 2023-03-13 | 2023-03-13 | Flight student conversation normalization scoring method and system based on voice recognition |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310236989.5A CN115938347B (en) | 2023-03-13 | 2023-03-13 | Flight student conversation normalization scoring method and system based on voice recognition |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115938347A CN115938347A (en) | 2023-04-07 |
CN115938347B true CN115938347B (en) | 2023-05-23 |
Family
ID=85818616
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310236989.5A Active CN115938347B (en) | 2023-03-13 | 2023-03-13 | Flight student conversation normalization scoring method and system based on voice recognition |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115938347B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117593924B (en) * | 2024-01-19 | 2024-03-26 | 中国民用航空飞行学院 | Scene reproduction-based air traffic controller training method and system |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109101518A (en) * | 2018-05-21 | 2018-12-28 | 全球能源互联网研究院有限公司 | Phonetic transcription text quality appraisal procedure, device, terminal and readable storage medium storing program for executing |
CN110335609A (en) * | 2019-06-26 | 2019-10-15 | 四川大学 | A kind of air-ground communicating data analysis method and system based on speech recognition |
CN110490428A (en) * | 2019-07-26 | 2019-11-22 | 合肥讯飞数码科技有限公司 | Job of air traffic control method for evaluating quality and relevant apparatus |
CN113066499A (en) * | 2021-03-12 | 2021-07-02 | 四川大学 | Method and device for identifying identity of land-air conversation speaker |
CN114049893A (en) * | 2022-01-11 | 2022-02-15 | 四川大学 | Method for monitoring and analyzing illegal operation of controller based on air traffic control voice recognition |
CN114187893A (en) * | 2021-12-15 | 2022-03-15 | 首都机场集团有限公司 | Intelligent voice monitoring method and system for air traffic control |
CN115223558A (en) * | 2022-06-24 | 2022-10-21 | 重庆中科云从科技有限公司 | Method, system and computer storage medium for managing air traffic control voice |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9218339B2 (en) * | 2011-11-29 | 2015-12-22 | Educational Testing Service | Computer-implemented systems and methods for content scoring of spoken responses |
US20130185314A1 (en) * | 2012-01-16 | 2013-07-18 | Microsoft Corporation | Generating scoring functions using transfer learning |
-
2023
- 2023-03-13 CN CN202310236989.5A patent/CN115938347B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109101518A (en) * | 2018-05-21 | 2018-12-28 | 全球能源互联网研究院有限公司 | Phonetic transcription text quality appraisal procedure, device, terminal and readable storage medium storing program for executing |
CN110335609A (en) * | 2019-06-26 | 2019-10-15 | 四川大学 | A kind of air-ground communicating data analysis method and system based on speech recognition |
CN110490428A (en) * | 2019-07-26 | 2019-11-22 | 合肥讯飞数码科技有限公司 | Job of air traffic control method for evaluating quality and relevant apparatus |
CN113066499A (en) * | 2021-03-12 | 2021-07-02 | 四川大学 | Method and device for identifying identity of land-air conversation speaker |
CN114187893A (en) * | 2021-12-15 | 2022-03-15 | 首都机场集团有限公司 | Intelligent voice monitoring method and system for air traffic control |
CN114049893A (en) * | 2022-01-11 | 2022-02-15 | 四川大学 | Method for monitoring and analyzing illegal operation of controller based on air traffic control voice recognition |
CN115223558A (en) * | 2022-06-24 | 2022-10-21 | 重庆中科云从科技有限公司 | Method, system and computer storage medium for managing air traffic control voice |
Also Published As
Publication number | Publication date |
---|---|
CN115938347A (en) | 2023-04-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP3432197B1 (en) | Method and device for identifying characters of claim settlement bill, server and storage medium | |
CN110704633B (en) | Named entity recognition method, named entity recognition device, named entity recognition computer equipment and named entity recognition storage medium | |
CN108734379B (en) | Online training method for realizing differentiation of customer service personnel | |
CN115938347B (en) | Flight student conversation normalization scoring method and system based on voice recognition | |
CN113468296B (en) | Model self-iteration type intelligent customer service quality inspection system and method capable of configuring business logic | |
CN112417132B (en) | New meaning identification method for screening negative samples by using guest information | |
CN111414751A (en) | Quality inspection optimization method, device, equipment and storage medium | |
CN114841293B (en) | Multimode data fusion analysis method and system for power Internet of things | |
CN112507376A (en) | Sensitive data detection method and device based on machine learning | |
CN115470796A (en) | Text instruction generation method and equipment for air traffic control simulation training | |
CN117768618A (en) | Method for analyzing personnel violation based on video image | |
CN116503042B (en) | Method and system suitable for processing training data of inspection and maintenance personnel | |
CN110010123A (en) | English phonetic word pronunciation learning evaluation system and method | |
CN112232078A (en) | Scheduling operation ticket auditing method based on bidirectional GRU and attention mechanism | |
CN116451646A (en) | Standard draft detection method, system, electronic equipment and storage medium | |
CN112992156B (en) | Power distribution network dispatching identity authentication system based on voiceprint authentication | |
CN113988064A (en) | Semi-automatic entity labeling monitoring method | |
CN110111810B (en) | Voice personality prediction method based on convolutional neural network | |
CN113822514A (en) | Full-media manuscript quality control method | |
CN112863493A (en) | Voice data labeling method and device and electronic equipment | |
CN112686238B (en) | Deep learning-based shipping bill identification method | |
CN116524926B (en) | Method for generating service form through voice control at mobile terminal | |
CN110096257B (en) | Design graph automatic evaluation system and method based on intelligent recognition | |
CN113722559A (en) | Training method, using method and related device of data demand recognition model | |
CN114818929A (en) | Self-learning labeling-based labeling model training method and labeling method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |