CN112862507A - Method, device, equipment, medium and product for preventing network appointment vehicle driver and passenger disputes - Google Patents
Method, device, equipment, medium and product for preventing network appointment vehicle driver and passenger disputes Download PDFInfo
- Publication number
- CN112862507A CN112862507A CN202110276262.0A CN202110276262A CN112862507A CN 112862507 A CN112862507 A CN 112862507A CN 202110276262 A CN202110276262 A CN 202110276262A CN 112862507 A CN112862507 A CN 112862507A
- Authority
- CN
- China
- Prior art keywords
- dispute
- driver
- car
- network
- passenger
- 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 65
- 238000012549 training Methods 0.000 claims abstract description 84
- 230000006399 behavior Effects 0.000 claims abstract description 34
- 239000004020 conductor Substances 0.000 claims abstract description 19
- 238000004590 computer program Methods 0.000 claims abstract description 13
- 239000013598 vector Substances 0.000 claims description 36
- 230000006870 function Effects 0.000 claims description 16
- 230000002265 prevention Effects 0.000 claims description 14
- 230000005764 inhibitory process Effects 0.000 claims description 5
- 238000004891 communication Methods 0.000 abstract description 33
- 238000004364 calculation method Methods 0.000 abstract description 13
- 238000005516 engineering process Methods 0.000 abstract description 11
- 238000012544 monitoring process Methods 0.000 abstract description 10
- 238000010801 machine learning Methods 0.000 description 22
- 230000006855 networking Effects 0.000 description 17
- 230000008569 process Effects 0.000 description 14
- 238000011161 development Methods 0.000 description 10
- 238000013461 design Methods 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Images
Classifications
-
- 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
- G06Q30/00—Commerce
- G06Q30/01—Customer relationship services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3343—Query execution using phonetics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/602—Providing cryptographic facilities or services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/62—Protecting access to data via a platform, e.g. using keys or access control rules
- G06F21/6218—Protecting access to data via a platform, e.g. using keys or access control rules to a system of files or objects, e.g. local or distributed file system or database
- G06F21/6245—Protecting personal data, e.g. for financial or medical purposes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/64—Protecting data integrity, e.g. using checksums, certificates or signatures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Bioethics (AREA)
- Computer Security & Cryptography (AREA)
- Computer Hardware Design (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Medical Informatics (AREA)
- Business, Economics & Management (AREA)
- Computing Systems (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Finance (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Acoustics & Sound (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Human Computer Interaction (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method, terminal equipment, storage medium and computer program product for preventing network car booking driver and passenger disputes, which collect real-time dialogue voice between network car booking drivers and passengers; inputting the real-time conversation voice into a preset driver and passenger dispute recognition model so that the driver and passenger dispute recognition model can recognize the real-time conversation voice and output a recognition result, wherein the driver and passenger dispute recognition model is obtained by carrying out federal learning training according to a network appointment vehicle driver and passenger conversation sample; and if the identification result identifies dispute behaviors among the drivers and conductors, outputting a preset dispute stop prompt to stop the dispute. The invention can only recognize whether the driver and the passenger have the dispute behavior through the real-time dialogue voice between the driver and the passenger, and outputs the prompt to timely stop the dispute behavior under the condition of recognizing that the dispute behavior exists, and does not need to rely on complicated video monitoring and other technologies, thereby reducing the requirements on terminal calculation capacity and communication bandwidth.
Description
Technical Field
The invention relates to the technical field of federal learning, in particular to a method and a device for preventing network appointment vehicle-driver disputes, terminal equipment, a storage medium and a computer program product.
Background
With the rapid development of technologies such as artificial intelligence and 5G communication, the intelligent car networking has gradually become the mainstream development trend in the production, manufacturing and use of cars, and under this environment, the application of the network car networking in the network car market is in the development period of rapid development.
In recent years, in the network car reservation market, frequent news events of driver-driver disputes and even driver-driver conflicts occur, which greatly hinders the development process of the network car reservation market. Based on various news reports, it is easy to find that the events such as driver and passenger disputes and driver and passenger conflicts are mostly caused by the problem of language communication between the driver and passengers of the vehicle. Although there are existing measures for applying a series of mature technologies such as voice or video monitoring to a vehicle-mounted terminal or an intelligent mobile terminal of a driver and a crew to perform real-time monitoring on behaviors such as communication and actions between the drivers and the crew, when applying the technologies such as voice or video monitoring, higher requirements are required for computing power of data processing, data transmission bandwidth and the like, not only is the data processing process complicated and response slow, but also the existing voice or video monitoring is only a simple record on the behaviors between the drivers and the crew, and cannot be prevented from language communication which may cause disputes or conflicts between the drivers and the crew.
In conclusion, the existing network car reservation market is based on the application of the intelligent car networking, so that the dispute or conflict of drivers and passengers cannot be prevented in time, and the effective development of the application of the intelligent car networking on the network car reservation market cannot be ensured.
Disclosure of Invention
The invention mainly aims to provide a method, terminal equipment, storage medium and computer program product for preventing a vehicle-to-vehicle and driver-to-driver dispute in a network appointment car market, and aims to solve the technical problem that the effective development of the vehicle-to-vehicle and driver-to-driver dispute or driver-to-driver conflict cannot be guaranteed due to the fact that the vehicle-to-driver dispute or driver-to-driver conflict is difficult to timely stop on the basis of intelligent vehicle network application in the network appointment car market in the prior network appointment car market.
In order to achieve the above object, the present invention provides a method for preventing a network car-booking and driver-passenger dispute, which is applied to a client, and comprises:
collecting real-time dialogue voice among the drivers and passengers of the online taxi appointment;
inputting the real-time conversation voice into a preset driver and passenger dispute recognition model so that the driver and passenger dispute recognition model can recognize the real-time conversation voice and output a recognition result, wherein the driver and passenger dispute recognition model is obtained by carrying out federal learning training according to a network appointment vehicle driver and passenger conversation sample;
and if the identification result identifies dispute behaviors among the drivers and conductors, outputting a preset dispute stop prompt to stop the dispute.
Further, the method for preventing vehicle-sharing disputes further includes:
and carrying out federal learning training according to the network appointment vehicle driving conversation sample to obtain a driving dispute recognition model.
Further, each client is connected with the server to form a learning federation, and the step of carrying out federated learning training according to the network appointment vehicle and driver-passenger conversation sample to obtain a driver-passenger dispute recognition model comprises the following steps:
encrypting and uploading a network appointment user identifier to the server so that the server can perform sample alignment on the network appointment car driving and taking conversation samples according to each encrypted network appointment user identifier;
receiving training samples which are sent by the server and correspond to the network appointment user identifications after sample alignment, and performing local model training by using the training samples to generate model parameters;
and uploading the model parameters to the server so that the server updates the local department dispute recognition models to be confirmed of the clients according to the model parameters until the department dispute recognition models converge or reach a preset iteration training round.
Further, the step of encrypting and uploading a car booking user identifier to the server so that the server performs sample alignment on the car booking driving and taking conversation samples according to each encrypted car booking user identifier includes:
encrypting the local network car booking user identification by using the secret key issued by the server to obtain an encrypted network car booking user identification;
and uploading the encrypted network car booking user identifications to the server so that the server decrypts the network car booking user identifications, determines the same network car booking user identification in the decrypted network car booking user identifications, and performs global sample alignment on the network car booking driver and passenger conversation samples according to the same network car booking user identification.
Further, the model parameters are uploaded after the client side encrypts by using a key, and the model parameters include: the step of uploading the model parameters to the server so that the server updates the local to-be-confirmed driver and crew dispute identification model of each client according to each model parameter includes:
encrypting the model parameters by using a secret key and uploading the model parameters to the server;
receiving the loss value fed back by the server, wherein the server decrypts the model parameter to obtain a decrypted network car-booking and car-riding conversation feature vector, and calls a preset loss function to calculate the network car-booking and car-riding conversation feature vector to obtain the loss value;
and locally calculating a gradient value according to the loss value so as to update a local to-be-confirmed department dispute identification model by using the gradient value.
Further, the dispute stop prompt is output in a voice and/or video broadcasting mode to prompt the driver and crew to execute dispute emergency operation, wherein the dispute emergency operation includes emergency video recording and/or alarming.
In order to achieve the above object, the present invention provides a network car-booking dispute inhibition device applied to a client, including:
the conversation acquisition module is used for acquiring real-time conversation voice among the drivers and passengers of the online appointment car;
the dispute recognition module is used for inputting the real-time conversation voice into a preset driver and passenger dispute recognition model so that the driver and passenger dispute recognition model can recognize the real-time conversation voice and output a recognition result, wherein the driver and passenger dispute recognition model is obtained by carrying out federal learning training according to a network appointment vehicle driver and passenger conversation sample;
and the dispute stopping module is used for outputting a preset dispute stopping prompt to stop disputes if dispute behaviors exist among the drivers and conductors identified by the identification result.
The steps of the method for preventing the network car-booking and driver-passenger dispute are realized when each functional module of the device for preventing the network car-booking and driver-passenger dispute operates.
In addition, to achieve the above object, the present invention also provides a terminal device, including: the system comprises a memory, a processor and a network car-booking and driver-driver dispute inhibition program which is stored in the memory and can run on the processor, wherein the steps of the network car-booking and driver-driver dispute inhibition method are realized when the processor executes the network car-booking and driver-driver dispute inhibition program.
In order to achieve the above object, the present invention further provides a storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method for suppressing vehicle-to-vehicle disputes as described above.
Furthermore, to achieve the above object, the present invention further provides a computer program product, which includes a computer program, and which, when being executed by a processor, implements the steps of the method for suppressing network appointment vehicle-driver disputes as described above.
The method, the device, the terminal equipment, the storage medium and the computer program product for preventing the network car booking driver and passenger disputes collect real-time dialogue voice among the network car booking drivers and passengers; inputting the real-time conversation voice into a preset driver and passenger dispute recognition model so that the driver and passenger dispute recognition model can recognize the real-time conversation voice and output a recognition result, wherein the driver and passenger dispute recognition model is obtained by carrying out federal learning training according to a network appointment vehicle driver and passenger conversation sample; and if the identification result identifies dispute behaviors among the drivers and conductors, outputting a preset dispute stop prompt to stop the dispute.
When the method is used for preventing potential disputes among the network car and the drivers, the real-time conversation voice among the network car and the drivers is collected through the client, then the real-time conversation voice is input into a driver and crew dispute recognition model obtained by carrying out federal learning training on the client and the server together according to a network car and driver conversation sample, so that the driver and crew dispute recognition model can recognize the real-time conversation voice and output a recognition result, and finally, if the client finds that the recognition result output by the driver and crew recognition model identifies that dispute behaviors do exist among the drivers and crew of the network car, the client immediately outputs a preset dispute prevention prompt so as to prevent the dispute behaviors in time.
Compared with the application of the intelligent car networking in the existing network car booking market, the intelligent car networking system can recognize whether dispute behaviors exist among drivers and passengers through real-time conversation voice among the drivers and passengers, and output prompts to timely stop the dispute under the condition that the dispute behaviors exist, does not need to rely on complicated video monitoring and other technologies, reduces the requirements on terminal calculation capacity and communication bandwidth, and ensures the effective development of the intelligent car networking in the network car booking market.
On the other hand, in the process of dispute recognition and timely prevention among the network car booking drivers and passengers, the dispute recognition model obtained based on the federal learning technology training is used, and the whole real-time conversation voice collection, dispute recognition and prevention operation are completed locally based on the client side, so that the relevant data cannot be leaked outwards, and the privacy and the safety of the voice data of the network car booking drivers and passengers are further protected.
Drawings
Fig. 1 is a schematic structural diagram of the hardware operation of a terminal device according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart illustrating an embodiment of a method for suppressing vehicle-sharing and driver-passenger disputes according to the present invention;
FIG. 3 is a schematic flow chart illustrating another embodiment of a method for stopping vehicle-sharing and driver-passenger disputes according to the present invention;
fig. 4 is a schematic view of an application flow involved in an embodiment of a method for stopping vehicle-sharing and driver-and-passenger disputes according to the present invention;
fig. 5 is a schematic block structure diagram of a network car booking driver-passenger dispute deterrence device.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
As shown in fig. 1, fig. 1 is a schematic structural diagram of a hardware operating environment related to a terminal device according to an embodiment of the present invention.
It should be noted that fig. 1 is a schematic structural diagram of a hardware operating environment of the terminal device. The terminal device in the embodiment of the invention is the client, and the client can be a mobile terminal, a vehicle-mounted terminal, a PC, a portable computer and other terminal devices.
As shown in fig. 1, the terminal device may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, a communication bus 1002. Wherein a communication bus 1002 is used to enable connective communication between these components. The user interface 1003 may include a Display screen (Display), an input unit such as a Keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1005 may be a high-speed RAM memory or a non-volatile memory (e.g., a magnetic disk memory). The memory 1005 may alternatively be a storage device separate from the processor 1001.
Those skilled in the art will appreciate that the terminal device configuration shown in fig. 1 is not intended to be limiting of the terminal device and may include more or fewer components than those shown, or some components may be combined, or a different arrangement of components.
As shown in fig. 1, a memory 1005, which is a kind of computer storage medium, may include therein an operating system, a network communication module, a user interface module, and a distributed task processing program. Among them, the operating system is a program that manages and controls the hardware and software resources of the sample terminal device, a handler that supports distributed tasks, and the execution of other software or programs.
In the terminal apparatus shown in fig. 1, the user interface 1003 is mainly used for data communication with each terminal; the network interface 1004 is mainly used for connecting a background server and performing data communication with the background server; and the processor 1001 may be configured to invoke a network appointment vehicle-to-vehicle dispute deterrence program stored in the memory 1005, and perform the following operations:
collecting real-time dialogue voice among the drivers and passengers of the online taxi appointment;
inputting the real-time conversation voice into a preset driver and passenger dispute recognition model so that the driver and passenger dispute recognition model can recognize the real-time conversation voice and output a recognition result, wherein the driver and passenger dispute recognition model is obtained by carrying out federal learning training according to a network appointment vehicle driver and passenger conversation sample;
and if the identification result identifies dispute behaviors among the drivers and conductors, outputting a preset dispute stop prompt to stop the dispute.
Further, the processor 1001 may call a network appointment vehicle-driver dispute deterrence program stored in the memory 1005, and further perform the following operations:
and carrying out federal learning training according to the network appointment vehicle driving conversation sample to obtain a driving dispute recognition model.
Further, each of the clients is connected with the server to form a learning federation, and the processor 1001 may call a network appointment vehicle-driver dispute stop program stored in the memory 1005, and further perform the following operations:
encrypting and uploading a network appointment user identifier to the server so that the server can perform sample alignment on the network appointment car driving and taking conversation samples according to each encrypted network appointment user identifier;
receiving training samples which are sent by the server and correspond to the network appointment user identifications after sample alignment, and performing local model training by using the training samples to generate model parameters;
and uploading the model parameters to the server so that the server updates the local department dispute recognition models to be confirmed of the clients according to the model parameters until the department dispute recognition models converge or reach a preset iteration training round.
Further, the processor 1001 may call a network appointment vehicle-driver dispute deterrence program stored in the memory 1005, and further perform the following operations:
encrypting the local network car booking user identification by using the secret key issued by the server to obtain an encrypted network car booking user identification;
and uploading the encrypted network car booking user identifications to the server so that the server decrypts the network car booking user identifications, determines the same network car booking user identification in the decrypted network car booking user identifications, and performs global sample alignment on the network car booking driver and passenger conversation samples according to the same network car booking user identification.
Further, the model parameters are uploaded after the client side encrypts by using a key, and the model parameters include: the network car-booking-driver conversation feature vector, and the processor 1001 may call the network car-booking-driver dispute stop program stored in the memory 1005, and further perform the following operations:
encrypting the model parameters by using a secret key and uploading the model parameters to the server;
receiving the loss value fed back by the server, wherein the server decrypts the model parameter to obtain a decrypted network car-booking and car-riding conversation feature vector, and calls a preset loss function to calculate the network car-booking and car-riding conversation feature vector to obtain the loss value;
and locally calculating a gradient value according to the loss value so as to update a local to-be-confirmed department dispute identification model by using the gradient value.
Further, the dispute stop prompt is output in a voice and/or video broadcasting mode to prompt the driver and the crew to execute dispute emergency operation, wherein the dispute emergency operation comprises emergency video recording and/or alarming.
Based on the above structure, the present invention provides various embodiments of the method for preventing vehicle-sharing and driver-riding disputes.
Referring to fig. 2, fig. 2 is a schematic flow chart of a method for preventing vehicle-sharing and driver-driver disputes according to a first embodiment of the present invention.
While a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than shown.
The method for preventing disputes among car owners of the network appointment car in the embodiments of the present invention is applied to a client for communicating with the internet outside the car on the network appointment car, and the client may be a mobile terminal, a vehicle-mounted terminal, a PC, a portable computer, or other terminal devices, and is not limited specifically herein.
The method for preventing the vehicle-booking and driver-and-passenger disputes in the embodiment comprises the following steps:
s100, collecting real-time dialogue voice among the drivers and passengers of the online taxi appointment;
the client side which is equipped with the intelligent car networking function on the network appointment car continuously collects real-time dialogue voice between a driver and a passenger (hereinafter referred to as a driver and a passenger for short) in the whole travel of the network appointment car.
It should be noted that, in this embodiment, the client may establish a connection with a communication network outside the network appointment car through a mature internet of things technology, so as to implement the function of the intelligent car networking. In addition, the client can continuously aim at real-time conversation voice among drivers and passengers in the online appointment car through a built-in or external microphone.
Specifically, for example, the client may be a certain vehicle-mounted terminal supporting a vehicle networking function on a car of the appointment, and the vehicle-mounted terminal is automatically turned on through information transmitted by sensors arranged on seats other than a current car driver seat of the appointment to collect real-time conversation voice generated among drivers and passengers in the car of the appointment.
It should be noted that, in this embodiment, the sensors disposed on the other seats may transmit corresponding information to the vehicle-mounted terminal when monitoring whether a passenger sits on the current seat, that is, when monitoring that the passenger sits on the current seat, the sensors immediately transmit a "passenger has got on the vehicle" information to the vehicle-mounted terminal, so that the vehicle-mounted terminal automatically starts to collect real-time conversation voice between drivers and passengers after receiving the information.
Further, in another possible embodiment, besides the vehicle-mounted terminal is automatically started, the vehicle-mounted terminal may be started based on receiving a command triggered by a driver or a passenger to collect real-time conversation voice between the drivers and the passengers, for example, after the driver of the networked car appointment starts to start the current formation by getting on the car, the vehicle-mounted terminal starts to continuously collect the real-time conversation voice generated between the drivers and the passengers in the networked car appointment by using a preset physical or virtual key on the vehicle-mounted terminal or by triggering and inputting a start command to the vehicle-mounted terminal based on a voice manner, so that the vehicle-mounted terminal starts to continuously collect the real-time conversation voice generated between the drivers and the passengers in the networked car appointment after receiving the start.
Step S200, inputting the real-time conversation voice into a preset driver and passenger dispute recognition model so that the driver and passenger dispute recognition model can recognize the real-time conversation voice and output a recognition result, wherein the driver and passenger dispute recognition model is obtained by carrying out federal learning training on a driver and passenger conversation sample according to a network appointment vehicle;
it should be noted that, in this embodiment, the client may combine with other clients and the server to form a federation, so that a federal learning training is performed on a network appointment car and driver-and-passenger conversation sample which is well collected and formulated based on the federation, so as to obtain a driver-and-passenger dispute recognition model, and thus disputes which may exist among drivers and passengers are recognized by using the driver-and-passenger dispute recognition model.
After acquiring real-time conversation voice among drivers and conductors, the client immediately takes the real-time conversation voice as model input and inputs the model input into a trained driver and conductor dispute recognition model, so that the driver and conductor dispute recognition model trains aiming at the model input to recognize the real-time conversation voice and output a recognition result capable of identifying whether dispute behaviors exist among the drivers and conductors.
Specifically, for example, the in-vehicle terminal establishes a connection with a trusted server together with other plural in-vehicle terminals, thereby forming a machine learning federation, and then the collected and formulated network appointment car driving conversation samples are used for carrying out federal learning training, so as to obtain a driving dispute recognition model which can recognize real-time conversation voice among drivers and passengers to accurately judge whether disputes exist among the drivers and passengers, therefore, the vehicle-mounted terminal can carry out the federal learning training in the whole journey of the network appointment car, after the real-time dialogue voice between the driver and the crew is acquired, the real-time dialogue voice is immediately input into the driver and crew dispute recognition model, and the driver and passenger dispute recognition model carries out model training calculation aiming at the real-time dialogue voice and outputs a recognition result which is obtained after the real-time dialogue voice is recognized and identifies whether disputes exist among the driver and passenger.
It should be noted that, in the present embodiment, the department dispute recognition model may be a speech recognition model mature in the industry, for example, the department dispute recognition model may specifically be DFCNN (deep full probabilistic neural network) and LAS (Listen Attend and speak, an LAS model established based on a neural network structure named "Listen Attend and speak"). It should be understood that, based on different design requirements of practical applications, in different possible embodiments, the department dispute recognition model may also be different from other mature speech recognition models listed herein, and the method for stopping the network car-booking department dispute according to the present invention is not limited to the specific type of the department dispute recognition model.
And S300, if the identification result identifies dispute behaviors among the drivers and conductors, outputting a preset dispute stop prompt to stop the dispute.
The client inputs the acquired real-time conversation voice into the driver and passenger dispute recognition model, so that after the driver and passenger dispute recognition model recognizes the real-time conversation voice and outputs a recognition result, if the client detects that dispute behaviors exist among drivers and passengers identified by the recognition result, the client immediately inputs a preset dispute stop prompt to stop the dispute in time.
It should be noted that, in this embodiment, in the process of obtaining the driver-crew dispute recognition model by the client performing federal learning training in association with the machine learning federation established by other clients and the server, the driver-crew conversation sample adopted by the client includes a sample label for indicating that the conversation between the drivers is "good communication", or "communication obstacle exists", even "dispute behavior exists", and when the client trains the driver-crew dispute recognition model, the driver-crew dispute recognition model also performs learning training on the feature vectors of the corresponding driver-crew conversation sample based on the different sample labels, so that the driver-crew dispute recognition model recognizes the real-time conversation voice between the drivers and outputs the recognition result corresponding to the sample label of "good communication", "communication obstacle exists", or "dispute behavior exists", and then the result identification of the communication state between the driver and the passenger can be identified.
Further, in this embodiment, the dispute stop prompt is output in a voice and/or video broadcast manner to prompt the driver and the crew to execute a dispute emergency operation, where the dispute emergency operation includes an emergency video and/or an alarm.
Specifically, for example, the vehicle-mounted terminal as the client performs federal learning training in combination with other vehicle-mounted terminals to obtain the driver and crew dispute recognition model, and the digital identifiers 1, 2 and 3 corresponding to "good communication", "communication obstacle exists" and "dispute behavior exists" are respectively used as recognition results which may be output after recognizing the real-time conversation voice, so that if the vehicle-mounted terminal currently inputs the collected real-time conversation voice among the drivers and crew into the driver and crew recognition model, and the driver and crew recognition model performs model training calculation based on the real-time conversation voice to output a recognition result of "3", the vehicle-mounted terminal determines that the communication among the drivers and crew in the current network appointment car journey is dispute through the sample label of "dispute behavior" corresponding to the recognition result of "3", and thus, the vehicle-mounted terminal immediately outputs to the drivers and crew in the network appointment car through a voice broadcast manner that the video recording is to be executed urgently and is about to be performed The alarm prompt can frighten the drivers and the passengers so as to effectively stop disputes caused by communication among the drivers and the passengers in time.
Further, in another feasible embodiment, after the client inputs the collected real-time dialogue speech into the driver and passenger dispute recognition model, so that the driver and passenger dispute recognition model recognizes the real-time dialogue speech and outputs a recognition result, if the client detects that disputes do not exist among drivers and passengers identified by the recognition result, the client continues to continuously collect the real-time dialogue speech, and inputs the real-time dialogue speech into the driver and passenger dispute recognition model to recognize and output the recognition result.
Specifically, for example, if the vehicle-mounted terminal inputs the collected real-time dialogue voice between the drivers and the conductors into the driver dispute recognition model, and the recognition result of the driver and passenger dispute recognition model which is output by model training calculation based on the real-time dialogue voice is '1' or '2', the vehicle-mounted terminal determines that no dispute exists between the drivers and the passengers in the current network appointment journey through the sample labels of good communication and communication obstacle corresponding to the identification result of 1 or 2 respectively, thus, the vehicle-mounted terminal continuously collects the real-time voice conversation among the drivers and the passengers in the journey of the network appointment and inputs the real-time voice conversation into the driver and passenger dispute recognition model, and the driver and passenger dispute recognition model carries out model training calculation based on the real-time dialogue voice and outputs a recognition result.
In the embodiment, the real-time conversation voice between a driver and a passenger (hereinafter referred to as a driver and a passenger) is continuously collected in the whole journey of the online appointment vehicle by a client which is equipped on the online appointment vehicle and realizes the intelligent vehicle networking function; after acquiring real-time conversation voice among drivers and conductors, a client immediately takes the real-time conversation voice as model input and inputs the model input into a trained driver and conductor dispute recognition model, so that the driver and conductor dispute recognition model trains aiming at the model input to recognize the real-time conversation voice and output a recognition result capable of identifying whether dispute behaviors exist among the drivers and conductors; the client inputs the acquired real-time conversation voice into the driver and passenger dispute recognition model, so that after the driver and passenger dispute recognition model recognizes the real-time conversation voice and outputs a recognition result, if the client detects that dispute behaviors exist among drivers and passengers identified by the recognition result, the client immediately inputs a preset dispute stop prompt to timely stop the dispute behaviors.
When the method is used for preventing potential disputes among the network car and the drivers, the real-time conversation voice among the network car and the drivers is collected through the client, then the real-time conversation voice is input into a driver and crew dispute recognition model obtained by carrying out federal learning training on the client and the server together according to a network car and driver conversation sample, so that the driver and crew dispute recognition model can recognize the real-time conversation voice and output a recognition result, and finally, if the client finds that the recognition result output by the driver and crew recognition model identifies that dispute behaviors do exist among the drivers and crew of the network car, the client immediately outputs a preset dispute prevention prompt so as to prevent the dispute behaviors in time.
Compared with the application of the intelligent car networking in the existing network car booking market, the intelligent car networking system can recognize whether dispute behaviors exist among drivers and passengers through real-time dialogue voice among the drivers and passengers, and output prompts to timely stop the dispute behaviors under the condition that dispute exists, does not need to rely on complicated video monitoring and other technologies, reduces the requirements on terminal calculation capacity and communication bandwidth, and ensures the effective development of the intelligent car networking in the network car booking market.
On the other hand, in the process of dispute recognition and timely prevention among the network car booking drivers and passengers, the dispute recognition model obtained based on the federal learning technology training is used, and the whole real-time conversation voice collection, dispute recognition and prevention operation are completed locally based on the client side, so that the relevant data cannot be leaked outwards, and the privacy and the safety of the voice data of the network car booking drivers and passengers are further protected.
Further, based on the first embodiment of the method for suppressing network car-booking and driver-passenger dispute according to the present invention, a second embodiment of the method for suppressing network car-booking and driver-passenger dispute according to the present invention is provided, and in this embodiment, the method for suppressing network car-booking and driver-passenger dispute according to the present invention may further include:
and step A, carrying out federal learning training according to the network appointment vehicle and driver conversation samples to obtain a driver and driver dispute recognition model.
Before the client specifically applies the driver-car dispute recognition model to recognize real-time conversation voice among the network appointment car drivers and passengers, a machine learning federation is constructed together with other clients, so that the built network appointment car driver-car dialogue samples collected in advance are utilized to carry out federal learning training based on the machine learning federation, and the driver-car dispute recognition model capable of accurately recognizing whether disputes exist among the network appointment car drivers and passengers is obtained.
It should be noted that, in this embodiment, the network appointment car driving and taking conversation sample may be specifically constructed based on the collected conversation voice between the drivers and the passengers in each journey of the network appointment car, and when constructing the corresponding network appointment car driving and taking conversation sample based on each conversation voice, the tags of "good communication", "communication obstacle exists", or "dispute exists" are further added based on the actual situation of the conversation voice of the drivers and the passengers or the background staff. Specifically, for example, after the dialogue voice of the driver and the passenger in one trip of the networked car appointment is collected, a prompt for selecting the dialogue voice tag in the current completed trip is output to the driver of the networked car appointment, so that the driver can select 'good communication' as the dialogue voice tag according to the actual situation that no communication problem occurs between the driver and the passenger in the current trip; or, if a communication obstacle or a dispute occurs between the driver and the passenger in the current trip, the driver correspondingly selects "communication obstacle exists" or "dispute exists" as the label of the conversation voice.
Further, in a possible embodiment, each of the clients is connected to the server to form a learning federation, please refer to the flow illustrated in fig. 3, where the step a may include:
step A01, encrypting and uploading a network car booking user identifier to the server, so that the server performs sample alignment on the network car booking driving and taking conversation samples according to each encrypted network car booking user identifier;
it should be noted that, in this embodiment, the current client and other clients each bound or associated with a corresponding network appointment form a machine learning federation together with a server respectively connected to and trusted by each client. It should be understood that, based on different design requirements of practical applications, in different possible embodiments, each client and even each server may be of the same type or different types of terminal devices, and the method for preventing vehicle-booking and driver-riding disputes according to the present invention is not limited to the specific types of the client and the server. In addition, in this embodiment, the online car booking user identifier may specifically be information such as a license plate number, a frame number, and/or a mobile phone number of a driver of the online car booking.
The client side encrypts the corresponding bound or associated network appointment vehicle user identification locally, and uploads the encrypted network appointment vehicle user identification to a server in a machine learning federation, so that the server can perform sample alignment operation on the constructed network appointment vehicle driving and taking conversation sample based on the network appointment vehicle user identification after receiving each encrypted network appointment vehicle user identification.
Further, in a possible embodiment, the step a01 may further include:
step A011, encrypting a local network car booking user identifier by using a secret key issued by the server to obtain an encrypted network car booking user identifier;
it should be noted that, in this embodiment, after each client and the server form the machine learning federation, the server distributes the key in the key pair to be generated to each client, so that each client performs encrypted data transmission with the server in the following. It should be understood that, based on different design requirements of practical applications, in different possible embodiments, the server may generate the key pair by using any mature key generation manner, and the method for preventing vehicle-department and driver-department disputes in the network is not limited by the process of generating the key pair by the server and the specific type of the key pair.
The client-side locally binds or associates the network car-booking user identification with the client-side, and the encrypted network car-booking user identification is obtained by encrypting the network car-booking user identification through a secret key issued by the server in advance.
And step A012, uploading the encrypted network car booking user identification to the server so that the server decrypts each network car booking user identification, determines the same network car booking user identification in each decrypted network car booking user identification, and performs global sample alignment on the network car booking driver and passenger conversation sample according to the same network car booking user identification.
The client side encrypts the bound or associated network car-booking user identifications by using the secret key issued by the server, and uploads the encrypted network car-booking user identifications to the server, so that the server decrypts the public key in the public key by using the pre-generated secret key after receiving the encrypted network car-booking user identifications, analyzes and contrasts whether the decrypted network car-booking user identifications are the same or not, thereby determining the same network car-booking user identification, and then the server performs global sample alignment operation on the constructed network car-booking driving conversation sample according to the same network car-booking user identification.
It should be noted that, in this embodiment, the server may specifically determine the same network car booking user identifier according to whether each decrypted network car booking user identifier belongs to a category. Specifically, for example, whether each network car booking user identifier is the same license plate number, the same frame number, the same mobile phone number, or the like is determined, so that a plurality of network car booking user identifiers belonging to the same license plate number, frame number, or mobile phone number are determined to be the same network car booking user identifier.
Step A02, receiving training samples which are sent by the server and correspond to the network appointment user identifications after sample alignment, and performing local model training by using the training samples to generate model parameters;
the client uploads the encrypted network appointment user identification to the server, so that the server decrypts the network appointment user identification and aligns global samples for network appointment car driving and taking conversation samples according to the same network appointment user identification, the server determines the network appointment car driving and taking conversation samples corresponding to the same network appointment user identification as training samples corresponding to the network appointment car user identification, then the server further issues the training samples to the corresponding client according to the same network appointment user identification, the client receives the training samples, and local model training is performed locally for an initial machine learning model to obtain model parameters of the machine learning model.
Specifically, for example, the machine learning model local to each client may specifically be a mature speech recognition model such as DFCNN, LAS, and the like, and the model parameters may specifically be network appointment, driver and passenger conversation feature vectors and sample labels output by the model. Referring to the application flow shown in fig. 4, after the client 1, the client 2, and the client 3 upload the network appointment user identifier (illustrated user ID)1, the network appointment user identifier 2, and the network appointment user identifier 3 encrypted by the key to the server, and the server decrypts the encrypted network appointment user identifier 1, the network appointment user identifier 2, and the network appointment user identifier 3 based on the public key corresponding to the key, it is determined that the network appointment user identifier 1 and the network appointment user identifier 2 are the same network appointment user identifiers belonging to the same category through analysis, the server performs a global sample alignment operation on the pre-constructed network appointment car riding conversation sample by using the network appointment user identifier 1 or the network appointment user identifier 2 together with the network appointment user identifier 3, and then the server further determines that the portion of the network appointment car riding conversation sample corresponding to the network appointment user identifier 1 or the network appointment user identifier 2 is the training sample 1, and determining the part of online appointment car driving and taking conversation samples corresponding to the online appointment car user identification 3 as training samples 2, respectively issuing the training samples 1 to the client 1 and the client 2, and issuing the training samples 2 to the client 3. Accordingly, after receiving the training sample 1 and the training sample 2, the client 1, the client 2, and the client 3 each start local model training on the initial machine learning model (DFCNN, LAS, or the like) locally, and then the client 1 trains the initial machine learning model locally using the training sample 1 to generate the car-to-car-riding conversation feature vector (illustrated car-to-car voice communication feature vector) 1 and the sample label 1, the client 2 trains the initial machine learning model locally using the training sample 1 to generate the car-to-car-riding conversation feature vector 2 and the sample label 2, and the client 3 trains the initial machine learning model locally using the training sample 2 to generate the car-to-car-riding conversation feature vector 3 and the sample label 3.
Step A03, uploading the model parameters to the server, so that the server updates the local department dispute recognition model to be confirmed at each client according to each model parameter until the department dispute recognition model converges or reaches a preset iteration training round.
The client-side carries out model training by using a corresponding training sample locally to obtain model parameters, then the model parameters are further uploaded to the server, so that the server carries out loss value calculation and feedback according to the model parameters, and updates the to-be-confirmed driver and conductor dispute recognition model which is trained locally by each client-side.
It should be noted that, in this embodiment, the preset iterative training round may be specifically generated by the client or the server based on the configuration of the staff, and is used to identify the number of times of model iterative training that has been completed by the client in the cyclic training update process for the department dispute recognition model in the current machine learning federation.
Further, in a possible embodiment, the model parameters are uploaded by the client after being encrypted by using a key, and step a03 may include:
step A031, upload to the said server after utilizing the cipher key to encrypt the said model parameter;
after the client performs model training locally by using the corresponding training sample to obtain the model parameters, the client further encrypts the model parameters by using a secret key issued by the server in advance, and uploads the encrypted model parameters to the server for loss value calculation.
Step A032, receiving the loss value fed back by the server, wherein the server decrypts the model parameter to obtain a decrypted network car-sharing and car-sharing conversation feature vector, and calls a preset loss function to calculate the network car-sharing and car-sharing conversation feature vector to obtain the loss value;
it should be noted that, in this embodiment, the preset loss value function may be any mature loss function, and based on different design requirements of practical applications, in different feasible implementation manners, the server may use different loss functions, so that the method for suppressing vehicle-sharing and driver-riding disputes in the present invention is not limited by the specific type of the preset loss value function.
After the client encrypts and uploads the model parameters obtained by local model training to the server, the server decrypts the encrypted model parameters through a local public key, then calls a preset loss value function to calculate a loss value, and feeds the loss value back to each client for model updating operation.
Specifically, for example, referring to the application flow shown in fig. 4, after the client 1 trains the initial machine learning model locally with the training sample 1 to generate the car-riding conversation feature vector 1 and the sample label 1, the client 2 trains the initial machine learning model locally with the training sample 1 to generate the car-riding conversation feature vector 2 and the sample label 2, and the client 3 trains the initial machine learning model locally with the training sample 2 to generate the car-riding conversation feature vector 3 and the sample label 3, the client 1, the client 2, and the client 3 encrypt and upload the car-riding conversation feature vector 1, the car-riding conversation feature vector 2, and the car-riding conversation feature vector 3 respectively with the key issued in advance by the server, and the server receives the encrypted car-riding conversation feature vector 1, the encrypted car-riding conversation feature vector 2, and the encrypted car-riding conversation feature vector 3, and receives the encrypted car-riding conversation feature vector 1, The network car-sharing and car-sharing conversation feature vector 2 and the network car-sharing and car-sharing conversation feature vector 3 are obtained through decryption by using a public key corresponding to the secret key, the network car-sharing and car-sharing conversation feature vector 1, the network car-sharing and car-sharing conversation feature vector 2 and the network car-sharing and car-sharing conversation feature vector 3 are obtained, then the server calls a loss function which is configured locally by a worker in advance, and loss values (graphic loss function values) are obtained through calculation based on the network car-sharing and car-sharing conversation feature vector 1, the network car-sharing and car-sharing conversation feature vector 2 and the network car-sharing and car-sharing conversation feature vector 3, and the loss values are encrypted and then fed back to the client 1, the client 2 and the client 3 respectively.
Step A033, calculating a gradient value locally according to the loss value, and updating a local to-be-confirmed driver-crew dispute identification model by using the gradient value.
After receiving the loss value fed back by the server, the client-side respectively calculates gradient values locally based on the loss value, so that the local to-be-confirmed driver and passenger dispute model is updated by the gradient values.
It should be noted that, in this embodiment, the client may specifically adopt any mature model gradient value calculation mode to calculate the gradient value based on the loss value issued by the server, and update the machine learning model that has been trained by the local model according to the gradient value. It should be understood that, based on different design requirements of practical applications, the calculation processes of calculating the gradient value by each client locally based on the loss value issued by the server may be different, and the method for preventing network appointment of vehicle and driver disputes of the present invention is not limited to the specific process of calculating the gradient by the client based on the loss value.
Further, in this embodiment, after the client locally updates the to-be-confirmed dispute recognition model (that is, the client locally uses the machine learning model that is model-trained by the training sample) based on the calculated gradient value, if the dispute recognition model is converged, or the number of times that the client currently uses the gradient value to update the dispute recognition model reaches the preset iteration training round, the client determines that the dispute recognition model is trained, so that the client can use the dispute recognition model to recognize the real-time dialogue speech between the network appointment car drivers and obtain an accurate recognition result of whether disputes exist between the drivers and the passengers.
In this embodiment, before the client specifically applies the department dispute recognition model to recognize the real-time conversation voice between the network appointment car drivers and passengers, a machine learning federation is first constructed together with other clients, so that the machine learning federation is used for carrying out federal learning training by using pre-collected constructed network appointment car driver and passenger conversation samples, and the department dispute recognition model capable of accurately recognizing whether disputes exist between the network appointment car drivers and passengers is obtained.
Therefore, when the method is used for preventing potential disputes among the network car and the drivers and the passengers, the real-time conversation voice between the network car and the drivers and the passengers collected by the client can be directly input into the driver and passenger dispute recognition model, so that the driver and passenger dispute recognition model can recognize the real-time conversation voice and output an accurate recognition result, and finally, the client can output a preset dispute prevention prompt when discovering that the recognition result identifies dispute behaviors among the drivers and passengers of the network car, thereby preventing the dispute behaviors in time.
Compared with the application of the intelligent car networking in the existing network car booking market, the intelligent car networking system and the method have the advantages that the dispute behavior of the drivers and the passengers is accurately identified by training the driver and passenger dispute identification model, so that the technology of complicated video monitoring and the like is not required, the requirements on terminal calculation capacity and communication bandwidth are reduced, and the effective development of the intelligent car networking in the network car booking market is ensured.
On the other hand, in the process of dispute recognition and timely prevention among the network car booking drivers and passengers, the dispute recognition model obtained based on the federal learning technology training is used, and the whole real-time conversation voice collection, dispute recognition and prevention operation are completed locally based on the client side, so that the relevant data cannot be leaked outwards, and the privacy and the safety of the voice data of the network car booking drivers and passengers are further protected.
In addition, referring to fig. 5, an embodiment of the present invention further provides a device for preventing vehicle-booking-driver dispute, where the device for preventing vehicle-booking-driver dispute is applied to a client, and the device includes:
the conversation acquisition module is used for acquiring real-time conversation voice among the drivers and passengers of the online appointment car;
the dispute recognition module is used for inputting the real-time conversation voice into a preset driver and passenger dispute recognition model so that the driver and passenger dispute recognition model can recognize the real-time conversation voice and output a recognition result, wherein the driver and passenger dispute recognition model is obtained by carrying out federal learning training according to a network appointment vehicle driver and passenger conversation sample;
and the dispute stopping module is used for outputting a preset dispute stopping prompt to stop disputes if dispute behaviors exist among the drivers and conductors identified by the identification result.
Preferably, the device for preventing disputes among car owners of the car appointment on the internet of the present invention further comprises:
and the model training module is used for carrying out federal learning training according to the network appointment vehicle and driver conversation samples to obtain a driver and driver dispute recognition model.
Preferably, each of the clients is connected to a server to form a learning federation, and the model training module includes:
the sample alignment unit is used for encrypting and uploading network appointment user identifications to the server so that the server can perform sample alignment on the network appointment car driving and taking conversation samples according to the encrypted network appointment user identifications;
the local training unit is used for receiving training samples which are sent by the server and correspond to the network appointment user identifications after sample alignment, and performing local model training by using the training samples to generate model parameters;
and the model updating unit is used for uploading the model parameters to the server so that the server updates the local department dispute recognition model to be confirmed at each client according to each model parameter until the department dispute recognition model converges or a preset iteration training round is reached.
Preferably, the sample alignment unit includes:
the first encryption transmission subunit is used for encrypting the local network car booking user identifier by using the secret key issued by the server to obtain an encrypted network car booking user identifier;
and the decryption alignment sample subunit is used for uploading the encrypted network car booking user identifications to the server so that the server can decrypt the network car booking user identifications, determine the same network car booking user identification in the decrypted network car booking user identifications, and perform global sample alignment on the network car booking driving and taking conversation samples according to the same network car booking user identification.
Preferably, the model parameters are uploaded after the client encrypts by using a key, and the model parameters include: the network appointment vehicle driver-riding dialogue feature vector, the model updating unit, also includes:
the second encryption transmission subunit is used for encrypting the model parameters by using the secret key and uploading the encrypted model parameters to the server;
the receiving subunit is configured to receive the loss value fed back by the server, where the server decrypts the model parameter to obtain a decrypted network car-sharing and car-sharing conversation feature vector, and calls a preset loss function to calculate the network car-sharing and car-sharing conversation feature vector to obtain the loss value;
and the calculating subunit is used for locally calculating a gradient value according to the loss value so as to update the local to-be-confirmed department dispute identification model by using the gradient value.
Preferably, the dispute stop prompt is output in a voice and/or video broadcasting manner to prompt the driver and the crew to execute a dispute emergency operation, wherein the dispute emergency operation includes emergency video recording and/or alarm.
The steps implemented by the functional modules of the device for preventing vehicle-sharing and driver-driver dispute in the invention during operation can refer to the embodiments of the method for preventing vehicle-sharing and driver-driver dispute in the invention, and are not described herein again.
In addition, an embodiment of the present invention further provides a terminal device, where the terminal device includes: the system comprises a memory, a processor and a network car-booking and driver-driver dispute prevention program which is stored in the memory and can run on the processor, wherein the steps of the network car-booking and driver-driver dispute prevention method are realized when the network car-booking and driver-driver dispute prevention program is executed by the processor.
The steps implemented when the network car-booking and driver-driver dispute deterrence program running on the processor is executed may refer to each embodiment of the network car-booking and driver-driver dispute deterrence method of the present invention, and are not described herein again.
In addition, an embodiment of the present invention further provides a storage medium applied to a computer, where the storage medium may be a non-volatile computer-readable storage medium, and the storage medium stores a network car-booking-driver-dispute deterrence program, and the network car-booking-driver-dispute deterrence program, when executed by a processor, implements the steps of the network car-booking-driver-dispute deterrence method described above.
In addition, the embodiment of the present invention further provides a computer program product, which includes an architecture program of store visitor information, and when executed by a processor, the architecture program of store visitor information implements the steps of the architecture method of store visitor information as described above.
The steps implemented when the network car-booking and driver-driver dispute deterrence program running on the processor is executed may refer to each embodiment of the network car-booking and driver-driver dispute deterrence method of the present invention, and are not described herein again.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal device (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present invention.
The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (10)
1. A method for stopping network car-booking and driver-passenger disputes is applied to a client side, and comprises the following steps:
collecting real-time dialogue voice among the drivers and passengers of the online taxi appointment;
inputting the real-time conversation voice into a preset driver and passenger dispute recognition model so that the driver and passenger dispute recognition model can recognize the real-time conversation voice and output a recognition result, wherein the driver and passenger dispute recognition model is obtained by carrying out federal learning training according to a network appointment vehicle driver and passenger conversation sample;
and if the identification result identifies dispute behaviors among the drivers and conductors, outputting a preset dispute stop prompt to stop the dispute.
2. The method for deterring vehicle-booking-driver dispute as claimed in claim 1, further comprising:
and carrying out federal learning training according to the network appointment vehicle driving conversation sample to obtain a driving dispute recognition model.
3. The network car-booking and dispute deterrence method according to claim 2, wherein the step of performing federal learning training according to the network car-booking and dispute dialogue sample to obtain a car-booking and dispute recognition model comprises the following steps:
encrypting and uploading a network appointment user identifier to the server so that the server can perform sample alignment on the network appointment car driving and taking conversation samples according to each encrypted network appointment user identifier;
receiving training samples which are sent by the server and correspond to the network appointment user identifications after sample alignment, and performing local model training by using the training samples to generate model parameters;
and uploading the model parameters to the server so that the server updates the local department dispute recognition models to be confirmed of the clients according to the model parameters until the department dispute recognition models converge or reach a preset iteration training round.
4. The method for deterring vehicle-sharing and driver-riding disputes according to claim 3, wherein said step of uploading encrypted vehicle-sharing user identifiers to said server for said server to perform sample alignment on said vehicle-sharing and driver-riding conversation samples according to each encrypted vehicle-sharing user identifier comprises:
encrypting the local network car booking user identification by using the secret key issued by the server to obtain an encrypted network car booking user identification;
and uploading the encrypted network car booking user identifications to the server so that the server decrypts the network car booking user identifications, determines the same network car booking user identification in the decrypted network car booking user identifications, and performs global sample alignment on the network car booking driver and passenger conversation samples according to the same network car booking user identification.
5. The method for stopping network appointment vehicle-driver-and-passenger disputes according to claim 3 or 4, wherein the model parameters are uploaded after the client side encrypts with a secret key, and the model parameters include: the step of uploading the model parameters to the server so that the server updates the local to-be-confirmed driver and crew dispute identification model of each client according to each model parameter includes:
encrypting the model parameters by using a secret key and uploading the model parameters to the server;
receiving the loss value fed back by the server, wherein the server decrypts the model parameter to obtain a decrypted network car-booking and car-riding conversation feature vector, and calls a preset loss function to calculate the network car-booking and car-riding conversation feature vector to obtain the loss value;
and locally calculating a gradient value according to the loss value so as to update a local to-be-confirmed department dispute identification model by using the gradient value.
6. The network appointment vehicle-driver-passenger dispute prevention method according to claim 1, wherein the dispute prevention prompt is a prompt which is output in a voice and/or video broadcasting manner to prompt the driver and passenger to execute dispute emergency operation, wherein the dispute emergency operation includes emergency video recording and/or alarm.
7. A network car-booking dispute inhibition device is applied to a client, and comprises:
the conversation acquisition module is used for acquiring real-time conversation voice among the drivers and passengers of the online appointment car;
the dispute recognition module is used for inputting the real-time conversation voice into a preset driver and passenger dispute recognition model so that the driver and passenger dispute recognition model can recognize the real-time conversation voice and output a recognition result, wherein the driver and passenger dispute recognition model is obtained by carrying out federal learning training according to a network appointment vehicle driver and passenger conversation sample;
and the dispute stopping module is used for outputting a preset dispute stopping prompt to stop disputes if dispute behaviors exist among the drivers and conductors identified by the identification result.
8. A terminal device, characterized in that the terminal device comprises: a memory, a processor and a network car-jockey dispute deterrence program stored on the memory and operable on the processor, the network car-jockey dispute deterrence program, when executed by the processor, implementing the steps of the network car-jockey dispute deterrence method as recited in any of claims 1-6.
9. A storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method of deterring network car-concession-driver-and-driver disputes according to any one of claims 1 to 6.
10. A computer program product comprising a computer program, wherein the computer program when executed by a processor implements the steps of a method of deterring network car-to-driver disputes as claimed in any one of claims 1 to 6.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110276262.0A CN112862507B (en) | 2021-03-15 | Method, device, equipment, medium and product for stopping network bus-tie disputes |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110276262.0A CN112862507B (en) | 2021-03-15 | Method, device, equipment, medium and product for stopping network bus-tie disputes |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112862507A true CN112862507A (en) | 2021-05-28 |
CN112862507B CN112862507B (en) | 2024-08-02 |
Family
ID=
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113409776A (en) * | 2021-06-30 | 2021-09-17 | 南京领行科技股份有限公司 | Voice recognition method and device, electronic equipment and storage medium |
CN113780850A (en) * | 2021-09-16 | 2021-12-10 | 广州宸祺出行科技有限公司 | Method and device for realizing network taxi appointment, driver and passenger guidance based on intelligent assistant |
CN114664001A (en) * | 2022-04-07 | 2022-06-24 | 乾三(北京)科技有限公司 | Method and device for acquiring driver and passenger information and processing driver and passenger dispute |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20040215449A1 (en) * | 2002-06-28 | 2004-10-28 | Philippe Roy | Multi-phoneme streamer and knowledge representation speech recognition system and method |
US20180342233A1 (en) * | 2017-05-23 | 2018-11-29 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for correcting speech recognition error based on artificial intelligence, and storage medium |
WO2020029585A1 (en) * | 2018-08-10 | 2020-02-13 | 深圳前海微众银行股份有限公司 | Neural network federation modeling method and device employing transfer learning, and storage medium |
CN111862953A (en) * | 2019-12-05 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Training method of voice recognition model, voice recognition method and device |
CN111933155A (en) * | 2020-09-18 | 2020-11-13 | 北京爱数智慧科技有限公司 | Voiceprint recognition model training method and device and computer system |
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20040215449A1 (en) * | 2002-06-28 | 2004-10-28 | Philippe Roy | Multi-phoneme streamer and knowledge representation speech recognition system and method |
US20180342233A1 (en) * | 2017-05-23 | 2018-11-29 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for correcting speech recognition error based on artificial intelligence, and storage medium |
WO2020029585A1 (en) * | 2018-08-10 | 2020-02-13 | 深圳前海微众银行股份有限公司 | Neural network federation modeling method and device employing transfer learning, and storage medium |
CN111862953A (en) * | 2019-12-05 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Training method of voice recognition model, voice recognition method and device |
CN111933155A (en) * | 2020-09-18 | 2020-11-13 | 北京爱数智慧科技有限公司 | Voiceprint recognition model training method and device and computer system |
Non-Patent Citations (1)
Title |
---|
黄蓉;: "福州市打车软件引争议", 福建质量管理, no. 2, pages 26 - 27 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113409776A (en) * | 2021-06-30 | 2021-09-17 | 南京领行科技股份有限公司 | Voice recognition method and device, electronic equipment and storage medium |
CN113409776B (en) * | 2021-06-30 | 2024-06-07 | 南京领行科技股份有限公司 | Voice recognition method and device, electronic equipment and storage medium |
CN113780850A (en) * | 2021-09-16 | 2021-12-10 | 广州宸祺出行科技有限公司 | Method and device for realizing network taxi appointment, driver and passenger guidance based on intelligent assistant |
CN114664001A (en) * | 2022-04-07 | 2022-06-24 | 乾三(北京)科技有限公司 | Method and device for acquiring driver and passenger information and processing driver and passenger dispute |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11335200B2 (en) | Method and system for providing artificial intelligence analytic (AIA) services using operator fingerprints and cloud data | |
US20210021687A1 (en) | Method and system for providing predictions via artificial intelligence (ai) models using a distributed system | |
EP3675121B1 (en) | Computer-implemented interaction with a user | |
US20180308344A1 (en) | Vehicle-to-infrastructure (v2i) accident management | |
CN109523652B (en) | Insurance processing method, device and equipment based on driving behaviors and storage medium | |
US20100305806A1 (en) | Portable Multi-Modal Emergency Situation Anomaly Detection and Response System | |
EP3496969A1 (en) | Method and system for providing information via collected and stored metadata using inferred attentional model | |
CN106817427A (en) | Car-mounted terminal and the system for car-mounted terminal | |
CN115225678A (en) | Driver identity online monitoring method and device and storage medium | |
DE102015104094A1 (en) | Telematics with variable reporting frequency | |
DE102015113644A1 (en) | DEVICE AND SYSTEM FOR PRODUCING VEHICLE DEVICE RECORDING DATA | |
DE102015107503A1 (en) | Method and system for starting an application | |
CN109726839A (en) | Net about vehicle monitoring method, apparatus and system | |
CN114417418A (en) | Privacy protection method and system for data processing of Internet of vehicles | |
CN112086098B (en) | Driver and passenger analysis method and device and computer readable storage medium | |
CN114879640A (en) | Automobile remote diagnosis system, method and computer readable storage medium | |
CN114407907A (en) | Intelligent driving system parameter self-adaptive adjusting method, equipment and storage medium | |
CN111862386A (en) | Accident recording method, device, medium and server for vehicle | |
CN115203078A (en) | Vehicle data acquisition system, method, equipment and medium based on SOA architecture | |
CN112862507A (en) | Method, device, equipment, medium and product for preventing network appointment vehicle driver and passenger disputes | |
CN109835280A (en) | A kind of system and vehicle showing vehicle-state and driving behavior by speech recognition | |
CN112862507B (en) | Method, device, equipment, medium and product for stopping network bus-tie disputes | |
US20230252790A1 (en) | Systems and methods for intelligent incident management in transportation environments | |
US20220414898A1 (en) | Fall prevention system | |
WO2018019015A1 (en) | Wireless charging method and device with intelligent monitoring |
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 |