WO2019037076A1 - 人工智能终端系统、服务器及其行为控制方法 - Google Patents

人工智能终端系统、服务器及其行为控制方法 Download PDF

Info

Publication number
WO2019037076A1
WO2019037076A1 PCT/CN2017/099015 CN2017099015W WO2019037076A1 WO 2019037076 A1 WO2019037076 A1 WO 2019037076A1 CN 2017099015 W CN2017099015 W CN 2017099015W WO 2019037076 A1 WO2019037076 A1 WO 2019037076A1
Authority
WO
WIPO (PCT)
Prior art keywords
behavior
data
artificial intelligence
server
intelligence terminal
Prior art date
Application number
PCT/CN2017/099015
Other languages
English (en)
French (fr)
Inventor
孙尚传
Original Assignee
深圳市得道健康管理有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 深圳市得道健康管理有限公司 filed Critical 深圳市得道健康管理有限公司
Priority to PCT/CN2017/099015 priority Critical patent/WO2019037076A1/zh
Priority to CN201780036358.1A priority patent/CN109313645B/zh
Publication of WO2019037076A1 publication Critical patent/WO2019037076A1/zh

Links

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/25—Integrating or interfacing systems involving database management systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/445—Program loading or initiating
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N7/00—Computing arrangements based on specific mathematical models
    • G06N7/02—Computing arrangements based on specific mathematical models using fuzzy logic

Definitions

  • the present invention relates to the field of computer technologies, and in particular, to an artificial intelligence terminal system, a server, and a behavior control method thereof.
  • Artificial intelligence is a new technical science that studies and develops theories, methods, techniques, and applications that simulate, extend, and extend human intelligence. Artificial intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that responds in a manner similar to human intelligence. Research in this area includes robotics, speech recognition, image recognition, Natural language processing and expert systems. Since the official introduction of the artificial intelligence discipline in 1956, over the past 50 years, it has made great progress and become a wide-ranging crossover and cutting-edge science. Up to now, the development of artificial intelligence has penetrated into many aspects of social life, liberating human beings from heavy physical strength, and at the same time gradually liberating human brain labor.
  • the inventors of the present invention have found that the existing intelligent terminals are still in the stage of weak artificial intelligence, and most of them still rely on the control of the user to perform corresponding behaviors.
  • a few intelligent terminals can perform some behaviors on their own according to external commands and environmental factors. However, these behaviors are sometimes unreasonable and do not meet the user's requirements for the terminal.
  • the technical problem to be solved by the embodiments of the present invention is to provide an artificial intelligence terminal system, a server, and a behavior control method thereof for using the behavior control data to standardize the behavior of the artificial intelligence terminal.
  • An embodiment of the present invention provides a method for controlling behavior of an artificial intelligence terminal, where the method includes: receiving behavior data of a first behavior to be executed uploaded by an artificial intelligence terminal; and performing behavior data of the first behavior to be executed and the established The behavior control data in the behavior control database is matched to obtain a first matching result; the first matching result is sent to the artificial intelligence terminal, so that the artificial intelligence terminal determines the artificial intelligence terminal according to the first matching result. Whether the first behavior can be performed.
  • an embodiment of the present invention further provides a server, including an input device, an output device, a memory, and a processor, where the memory stores behavior control data, and the processor is configured to: receive artificial intelligence by using the input device. Behavior data of the first behavior uploaded by the terminal; matching the behavior data of the first behavior to be executed with the behavior control data in the established behavior control database to obtain a first matching result; using the output device to A matching result is sent to the artificial intelligence terminal, so that the artificial intelligence terminal determines whether the artificial intelligence terminal can perform the first behavior according to the first matching result.
  • a server including an input device, an output device, a memory, and a processor, where the memory stores behavior control data, and the processor is configured to: receive artificial intelligence by using the input device. Behavior data of the first behavior uploaded by the terminal; matching the behavior data of the first behavior to be executed with the behavior control data in the established behavior control database to obtain a first matching result; using the output device to A matching result is sent to the artificial intelligence terminal, so that the artificial intelligence terminal determines whether the artificial intelligence terminal can
  • an embodiment of the present invention further provides an artificial intelligence control system, where the system includes a server as described above and an artificial intelligence terminal communicatively coupled to the server.
  • an embodiment of the present invention further provides a computer readable storage medium, where the computer storage medium stores a program, and the program can be executed to implement the behavior control method as described above.
  • the behavior control data may be used to standardize the behavior of the terminal, so that the terminal is The operation or action is more reasonable and meets the requirements of the user for the terminal.
  • FIG. 1 is a schematic block diagram showing the structure of an artificial intelligence terminal according to an embodiment of the present invention.
  • FIG. 2 is a schematic block diagram showing the structure of an artificial intelligence control system in an embodiment of the present invention.
  • FIG. 3 is a first schematic flowchart of a behavior control method according to an embodiment of the present invention.
  • FIG. 4 is a second schematic flowchart of a behavior control method according to an embodiment of the present invention.
  • FIG. 5 is a schematic block diagram showing the structure of a server in an embodiment of the present invention.
  • FIG. 6 is a third schematic flowchart of a behavior control method according to an embodiment of the present invention.
  • FIG. 7 is a schematic block diagram showing the structure of a behavior processing device in an embodiment of the present invention.
  • FIG. 8 is a first schematic flowchart of a method for establishing a behavior control database according to an embodiment of the present invention.
  • FIG. 9 is a second schematic flowchart of a method for establishing a behavior control database in an embodiment of the present invention.
  • FIG. 10 is a partial flow chart of a method for establishing a behavior control database in an embodiment of the present invention.
  • the embodiment of the invention provides an artificial intelligence terminal and a corresponding behavior control method, which are specifically described below:
  • an embodiment of the present invention provides an artificial intelligence terminal 100.
  • the smart terminal may include an input/output device 101, a processor 102, and a memory 103.
  • the input/output device 101 and the memory 103 are respectively connected to the processor 102.
  • a bus connection can be used.
  • the memory 103 stores therein behavior control data. It is to be understood that the artificial intelligence terminal 100 may not be provided with the input/output device 101, which is not limited herein.
  • the memory 103 can also store program code, and the processor can call the program code and execute the following method:
  • the smart terminal 100 may further include a sensor that can be used to collect data from the environment in real time; for example, the sensor may include an image sensor, an acoustic sensor, an odor sensor, a temperature sensor, a humidity sensor, a distance sensor, a position sensor, or the like. Multiple kinds of environment data are collected; after the environmental data is collected, the processor can process the environmental data to generate behavior data of the first behavior to be executed to respond to the surrounding environment;
  • an instruction input by a user or an instruction sent by another device such as a server is received, and the behavior data of the first behavior is obtained according to the instruction.
  • the behavior control data may be data established in advance according to an artificial intelligence ethics system based on altruism
  • the first behavior may be a combination of any one or more of operations, actions, calculations, inputs, outputs, and the like to be performed by the smart terminal;
  • the behavior control data may be data pre-stored in the memory 103, which may include a plurality of behaviors and corresponding executable degrees.
  • the plurality of behaviors may be divided into several categories, for example, according to Confucianism, interpretation, Tao, medicine, Wu, music, customs, science, business... are classified and stored separately.
  • the first matching result may be explicitly a match (executable) or a non-match (not executable).
  • the processor 102 determines whether there is executable behavior control data that matches the behavior data of the first behavior in the pre-stored behavior control data; if the result is a match, the control performs the first behavior; otherwise, the mismatch, Then the control does not perform the first action.
  • the first matching result can include an enforceability.
  • the processor 102 matches the behavior data of the first behavior with the pre-stored behavior control data to obtain a first executable degree of the first behavior; and determines, according to the first executable degree, whether the artificial intelligence terminal is executable. Said the first act. Furthermore, the processor 102 can control whether the artificial intelligence terminal performs the first behavior according to the determination result.
  • the processor 102 will use the behavior control data to match the behavior data of the behavior, and judge whether the behavior is executable according to the executable degree obtained by the matching, and the specific judgment method may Yes: A threshold is set in advance. If the degree of executableness is higher than the threshold, it is judged to be executable. Otherwise, it is determined to be unexecutable.
  • the specific value of the threshold can be set according to the empirical value and can be dynamically adjusted.
  • the intelligent terminal provided in the embodiment can store the behavior control data in the terminal in advance, and can use the behavior control data to standardize the operation or action of the terminal, so that the operation or action of the terminal is more reasonable, and the requirements of the user for the terminal are satisfied.
  • the smart terminal 100 can be a mobile phone, a computer, an intelligent robot, or other terminal with data analysis processing capability.
  • an intelligent robot any device that can be used to analyze data.
  • the behavior control data is stored in the memory of the intelligent robot.
  • the behavior control data is read-only data solidified in the robot memory, which prevents the behavior control data from being maliciously modified; in another embodiment Intelligent robots can have self-learning functions and use self-learning to update and correct behavior control data.
  • intelligent robots can be classified into different categories.
  • home robots can be equipped with functions such as moving, grabbing/handling items, and voice dialogue.
  • Industrial robots can be configured with different functions according to different purposes. For example, welding, handling, painting, grinding, palletizing, assembly, etc., military robots can have complex terrain movement, load-bearing, mine-clearing, shooting and other functions.
  • All kinds of intelligent robots can perform a variety of heavy work that humans are unable to do. However, in the course of working, robots often face some complicated or unexpected situations.
  • the processing method is that the robot stops working and needs instructions from the human user. This kind of processing is more secure, but obviously it is likely to cause delay in the robot work and waste the processing power of the robot; the second processing method is to give the robot a certain autonomic processing authority. , can react to the complex or unexpected situations it faces, perform appropriate operations or actions to deal with the situation, but bring certain risks: the robot's handling of complex or unexpected situations may not be consistent with the expectations of human users. It may even have extremely negative consequences.
  • the behavior control data may be pre-stored in the memory of the intelligent robot, and the related data of the processing scheme (ie, the behavior to be performed) generated by the robot in the face of complicated or unexpected situations is required.
  • the behavior control data After matching with the behavior control data, only the matching scheme can be executed, or further, only the highly executable scheme can be executed, and the behavior control data is defined by the human user, and the behavior that can be considered reasonable by humans Giving high enforceability, giving low degree of enforceability to unreasonable behavior, as long as the content of the behavior control data is reasonable, it can control the processing strategy of intelligent robots with certain autonomous processing capabilities in the face of complex or unexpected situations, preventing The negative consequences of the robot's unreasonable response.
  • an embodiment of the present invention further provides an artificial intelligence control system, which includes the artificial intelligence terminal 100 and the server 200 described above, and the server 200 communicates with the artificial intelligence terminal 100 in a wired or wireless manner.
  • the server 200 can be deployed in the cloud, where a behavior control database is stored, and the behavior control database includes a large amount of behavior control data, and can be continuously updated.
  • the input and output device 101 can be used to upload the behavior data of the first behavior to be executed by the terminal to the server 200, and the server 200 can match the behavior data of the first behavior with the behavior control database to obtain the second behavior.
  • the behavior control data stored in the memory of the robot may have a limited amount of data, and the matching result of the behavior may not be accurate.
  • the embodiment of the present invention further provides a solution, so that the robot can upload the generated processing scheme to the server, and then Receive the executableness of the server feedback for a more accurate judgment. Since the server can store a large amount of behavior control data, and it is constantly supplemented and revised, the server's judgment on behavior can be considered more reliable and accurate, so that the robot can judge more accurately by the executable degree of server feedback. Whether the generated processing scheme is executable. In view of this, the robot can assign a higher weight to the second executable degree when performing weighted averaging on the first executable degree and the second executable degree of the server feedback.
  • the embodiment of the present invention further provides a behavior control method, which can be applied to the foregoing artificial intelligence terminal 100, and can be specifically executed by the processor 102 of the artificial intelligence terminal 100.
  • the method can include:
  • the first behavior may be any operation or action to be performed by the smart terminal 100;
  • the behavior control data may be data pre-stored in the memory, where a plurality of behaviors may be included.
  • the plurality of behaviors may be divided into several categories and stored separately; further, the behavior control data may further include The executables corresponding to each of the behaviors it contains.
  • the first matching result includes the first executable degree of the first behavior, determining whether the first behavior can be performed according to the first executable degree.
  • the behavior control data is used to match the behavior data of the behavior, and the behavior is determined according to the executable degree obtained by the matching.
  • the specific determination method may be: preset A threshold value is judged to be executable if the degree of executableness is higher than the threshold value, and is determined to be unexecutable.
  • the specific value of the threshold can be set according to the empirical value and can be dynamically adjusted.
  • the behavior processing method provided in this embodiment performs the matching of the behavior executable degree through the pre-stored behavior control data, and can use the behavior control data to standardize the operation or action of the terminal, so that the operation or action of the terminal is more reasonable. Meet the user's requirements for the terminal.
  • the artificial intelligence terminal may use the feature extraction and feature matching manner to obtain the first executable degree of the first behavior.
  • the foregoing S302 may include:
  • the behavior data of the first behavior is parsed to extract behavior characteristic data of the first behavior from the behavior data.
  • the behavior feature may include at least one of the following characteristics: behavior time, behavior place, behavior subject, and behavior object.
  • the foregoing 302B may specifically include: in the behavior control data, the behavior data of the at least one behavior including each behavior characteristic of the first behavior, and the first executable of the first behavior according to the executable degree of the at least one behavior. degree. Further, the obtaining, according to the executable degree of the at least one behavior, the first executable degree of the first behavior comprises: if the at least one behavior is an behavior, the executable degree of the behavior is a first executable degree of the first behavior; if the at least one behavior includes a plurality of behaviors, weighting the plurality of executables of the plurality of behaviors to obtain a first executable degree of the first behavior.
  • the method may further include:
  • the server can be deployed in the cloud, which stores a behavior control database.
  • the behavior control database includes massive behavior control data and can be continuously updated.
  • the behavior control database in the server is the same as the behavior control database of the artificial intelligence terminal, and is the data established by the artificial intelligence moral system based on the principle of altruism.
  • the type of behavior in the behavior control database in the server can be set to be richer than the behavior in the behavior control data of the artificial intelligence terminal.
  • the manner in which the server obtains the second matching degree is similar to the manner in which the first matching result is obtained, and the second matching result may be explicitly matched (executable) or not matched (unexecutable).
  • the behavior control data in the server may be read-only data pre-stored in a server executing the method; or readable and writable data pre-stored in a server executing the method, but all of the artificial intelligence terminals There is no write permission; or it is readable and writable data stored in advance in the server executing the method, but only the artificial intelligence terminal that satisfies the set condition has write authority.
  • the foregoing 303 may include controlling execution of the first behavior according to the first matching result and the second matching result. For example, it is determined whether the first behavior can be performed according to the first executable degree and the second executable degree.
  • the foregoing 303 may include:
  • the first behavior cannot be performed.
  • the foregoing 301 can include:
  • 301B1 Process data collected by the sensor to generate behavior data of the first behavior to be performed.
  • the smart terminal 100 may further include a sensor that can be used to collect data from the environment in real time; for example, the sensor may include one or more of an image sensor, an acoustic sensor, an odor sensor, a temperature sensor, a humidity sensor, a distance sensor, a position sensor, and the like. Therefore, various environmental data are collected; after the environmental data is collected, the processor can process the environmental data to generate behavior data of the first behavior to be executed to respond to the surrounding environment.
  • a sensor that can be used to collect data from the environment in real time
  • the sensor may include one or more of an image sensor, an acoustic sensor, an odor sensor, a temperature sensor, a humidity sensor, a distance sensor, a position sensor, and the like. Therefore, various environmental data are collected; after the environmental data is collected, the processor can process the environmental data to generate behavior data of the first behavior to be executed to respond to the surrounding environment.
  • an embodiment of the present invention further provides a computer readable storage medium storing a program, the program being executable to implement the behavior processing method as described above.
  • the smart terminal 100 that performs the foregoing behavior processing method may be a mobile phone, a computer, an intelligent robot, or another terminal with data analysis processing capability.
  • an intelligent robot as an example:
  • the behavior control data is stored in the memory of the intelligent robot.
  • the behavior control data is read-only data that is solidified in the robot memory, and the behavior control data is prevented from being maliciously modified.
  • intelligent robots can be classified into different categories.
  • home robots can be equipped with functions such as moving, grabbing/handling items, and voice dialogue.
  • Industrial robots can be configured with different functions according to different purposes. For example, welding, handling, painting, grinding, palletizing, assembly, etc., military robots can have complex terrain movement, load-bearing, mine-clearing, shooting and other functions.
  • All kinds of intelligent robots can perform a variety of heavy work that humans are unable to do. However, in the course of working, robots often face some complicated or unexpected situations.
  • the processing method is that the robot stops working and needs instructions from the human user. This kind of processing is more secure, but obviously it is likely to cause delay in the robot work and waste the processing power of the robot; the second processing method is to give the robot a certain autonomic processing authority. , can react to the complex or unexpected situations it faces, perform appropriate operations or actions to deal with the situation, but bring certain risks: the robot's handling of complex or unexpected situations may not be consistent with the expectations of human users. It may even have extremely negative consequences.
  • the intelligent robot pre-stores the behavior control data, and when the robot faces a complex or unexpected situation, the related data of the generated processing scheme (that is, the behavior to be performed) needs to go through and control the behavior.
  • Data is matched, only matching schemes can be executed, or further, only highly executable schemes can be executed, and behavior control data is defined by human users, which can give high executableness to behaviors that humans think is reasonable. Degree, for the unreasonable behavior to give low enforceability, as long as the content of the behavior control data is reasonable, it can control the intelligent robot with certain autonomous processing ability in the face of complex or unexpected situations, to prevent the robot from unreasonable response The negative consequences.
  • the behavior control data stored in the memory of the robot may have a limited amount of data, and the matching result of the behavior may not be accurate.
  • the embodiment of the present invention further provides a solution, so that the robot can upload the generated processing scheme to the server, and then Receive the executableness of the server feedback for a more accurate judgment. Since the server can store a large amount of behavior control data, and it is constantly supplemented and revised, the server's judgment on behavior can be considered more reliable and accurate, so that the robot can judge more accurately by the executable degree of server feedback. Whether the generated processing scheme is executable. In view of this, the robot can assign a higher weight to the second executable degree when performing weighted averaging on the first executable degree and the second executable degree of the server feedback.
  • the embodiment of the present invention further provides a server 200, such as the foregoing server, connected to the foregoing artificial intelligence terminal, and can be used to manage the artificial intelligence terminal.
  • the server 200 includes an input/output device 201, a processor 202, and a memory 203.
  • the input/output device 201 and the memory 203 are respectively connected to the processor 202, and specifically, a bus connection method can be adopted.
  • the input/output device 101 can include means for receiving information, such as a receiver, a transmitter, and a touch screen, buttons, data interfaces such as a USB interface, a display screen, and the like.
  • the memory 203 stores behavior control data.
  • Program code may also be stored in the memory 203. As shown in FIG. 6, the processor 202 can call the program code and execute the following method:
  • the processor 104 receives the behavior data of the first behavior uploaded by the artificial intelligence terminal through the input/output device 101.
  • the artificial intelligence terminal can obtain the behavior data of the first behavior as described in the previous embodiment.
  • the processor 202 sends a behavior instruction to the artificial intelligence terminal through the input/output device 601 to cause the artificial intelligence terminal to parse the behavior instruction to obtain behavior data of the first behavior.
  • the behavior control database may be a database established in advance according to the artificial intelligence ethics system based on the principle of altruism.
  • the behavior control data may be data pre-stored in the memory 104, which may include multiple behaviors, and may also include corresponding executable degrees.
  • multiple behaviors may be divided into several categories, for example, according to Confucianism. , release, Tao, medicine, Wu, music, customs, science, business... separate classification for storage.
  • the first matching result may be explicitly a match (executable) or a non-match (not executable).
  • the processor 202 determines whether there is executable behavior control data matching the behavior data of the first behavior in the pre-stored behavior control data; wherein if the result is a match, the first behavior can be performed; otherwise If it does not match, it means that the first behavior cannot be performed.
  • the first matching result can include an enforceability.
  • the processor 202 matches the behavior data of the first behavior with the pre-stored behavior control data to obtain a first executable degree of the first behavior.
  • the artificial intelligence terminal may determine that the first behavior is performed as described in the previous embodiment.
  • the server for managing the artificial intelligence terminal provided in this embodiment can use the behavior control data to standardize the operation or action of the terminal by storing the behavior control data in the server in advance, so that the operation or action of the terminal is more reasonable and satisfied. User requirements for the terminal.
  • the behavior control data may be read-only data stored in the memory 203, and the behavior control data may be prevented from being maliciously modified.
  • the behavior control data is readable and writable data pre-stored in the memory 203, but all the artificial intelligence terminals connected to the server do not have write permission, and the server can still have modification authority, thereby preventing The behavior control data is maliciously modified by the artificial intelligence terminal, and the data can be updated.
  • the behavior control data is readable and writable data stored in advance in the execution memory 203, but only the artificial intelligence terminal that satisfies the set condition has write permission, and the server and the authenticated partial artificial intelligence terminal may still have Modifying the permissions, thereby avoiding malicious modification of the behavior control data by the unauthenticated artificial intelligence terminal, and also ensuring that the data is updatable.
  • the setting condition may be set according to actual requirements, for example, the identity of the terminal user has been verified, and the terminal level is the security level.
  • the server's memory 203 stores behavior control data.
  • the behavior control data is read-only data that is solidified in the server memory 203, which prevents the behavior control data from being maliciously modified;
  • the server may have a self-learning function and use the self-learning function to update and correct the behavior control data by itself.
  • the server 200 may use the feature extraction and feature matching manner to obtain the first matching result (such as the first executable degree of the first behavior).
  • the foregoing 602 performed by the processor 202 may include:
  • 602B Perform feature matching on the behavior feature of the first behavior with the behavior control data in the behavior control database, to obtain the first matching result.
  • the behavior control database can store executables with different behaviors and their behavioral characteristics, and different behaviors.
  • the enforceability of the behavior stored in the behavior control database may be self-learned by the processor 503 based on feedback from the user on the historical behavior of the terminal, or may be obtained by user input.
  • the 602B performed by the processor 503 may include:
  • 602B2 Obtain a first executable degree of the first behavior according to the executable degree of the at least one second behavior.
  • the 602B2 may be specifically: if the at least one second behavior has only one second behavior, the executable degree of the second behavior is directly used as the first executable degree of obtaining the first behavior;
  • the at least one second behavior includes a plurality of second behaviors, and the plurality of executables of the plurality of second behaviors are weighted and averaged to obtain a first executable degree of the first behavior.
  • the behavior feature may include at least one of the following characteristics: behavior time, behavior place, behavior subject, and behavior object.
  • the artificial intelligence terminal can match the behavior data of the first behavior to be executed with the behavior control data stored by itself according to the previous embodiment, and obtain the pre-judgment execution degree, wherein the pre-judgment execution degree is similar to the acquisition process.
  • the artificial intelligence terminal directly determines that the first behavior is unexecutable when the pre-judging execution degree is less than the first predetermined threshold, otherwise uploading the behavior data of the first behavior to be executed by the terminal to
  • the server 200 matches the behavior data of the first behavior with the stored behavior control database to obtain the actual executable degree of the behavior, and feeds the actual executable degree to the terminal; after receiving the actual executable degree, the terminal receives the actual executable degree.
  • the first predetermined threshold and the second predetermined threshold may be set according to actual conditions, and the numerical relationship may be set such that the first predetermined threshold is smaller than the second predetermined threshold.
  • the behavior control data stored in the memory of the artificial intelligence terminal may have a limited amount of data, the matching result of the behavior may not be accurate enough, so the behavior data of the first behavior after the terminal is judged may be uploaded to the server, and then Receive the executableness of the server feedback for a more accurate judgment.
  • the server can store a large amount of behavior control data, and it is constantly supplemented and revised, the server's judgment on behavior can be considered more reliable and accurate, so that the robot can judge more accurately by the executable degree of server feedback. Whether the generated processing scheme is executable. In view of this, the robot can assign a higher weight to the second executable degree when performing weighted averaging on the first executable degree and the second executable degree of the server feedback.
  • the embodiment of the invention further provides a behavior processing device and a corresponding establishment method, which are specifically described below:
  • a behavior processing device 700 may be included, and the behavior processing device 700 may include an input/output device 701, a processor 702, and a memory 703.
  • the input/output device 701 and the memory 703 are all connected to the processor 702, and specifically, a bus connection manner can be adopted.
  • the memory 703 stores a behavior control database for storing the foregoing behavior control data for providing to the artificial intelligence terminal or the server connected to the artificial intelligence terminal to obtain the executable degree of the execution behavior of the artificial intelligence terminal. Further, the artificial intelligence terminal determines whether to execute the to-be-executed behavior according to the acquired executable degree.
  • Program code may also be stored in the memory 703. As shown in FIG. 8, the processor 702 can call the program code and execute the following method:
  • the processor 703 obtains the behavior data of the first behavior by using the input/output device 701.
  • the input/output device 701 is any device that can obtain data from the outside world, such as a human-machine interaction interface, a remote control device, a microphone, a receiver, a collector, etc., for acquiring behavior data from the outside.
  • the processor 703 directly obtains behavior data input by the user through a human-machine interaction interface, a remote control device, or a microphone, or receives behavior data sent by other devices such as other artificial intelligence terminals or servers, or acquires acquired data such as an image collector. Behavioral data.
  • the first behavior may be a combination of any one or more of operations, actions, calculations, inputs, outputs, and the like executable by the artificial intelligence terminal.
  • the behavior feature may include at least one of the following characteristics: behavior time, behavior place, behavior subject, and behavior object.
  • each behavioral characteristic of the first behavior may be evaluated according to an artificial intelligence ethics system based on altruism, thereby obtaining a first executable degree of the first behavior.
  • the evaluation can also be continued in accordance with other set evaluation principles.
  • the artificial intelligence ethics system can be based on the "social honor and disgrace view”, “disciple rules”, “The Analects of Confucius”, “Tao Jing” and other altruistic behavioral rules as the data system.
  • the artificial intelligence moral system can contain different categories of moral content such as Confucianism, Islam, Taoism, Medicine, Wu, Music, Vulgar, Science, and Business.
  • the processor 702 matches the behavioral characteristics of the first behavior with the moral content of the artificial intelligence moral system, and the degree of compliance of the first behavior is obtained by the matching degree.
  • weights may be set for different behavioral features in advance, and the degree of matching of each behavioral feature is weighted and averaged to obtain the executable degree of the first behavior.
  • the behaviors stored in the behavior control database can be divided into several categories.
  • the first behavior can be classified according to Confucianism, Interpretation, Taoism, Medicine, Wu, Music, Vulgar, Science, and Business, respectively. Then store it.
  • the classification of the first behavior can be performed directly during the performance evaluation process of execution 803.
  • the data in the behavior control database can be stored as read-only data to avoid malicious modification of devices (such as artificial intelligence terminals or servers) that subsequently use the behavior control data.
  • the data in the behavior control database is stored as readable and writable data, but all artificial intelligence terminals do not have write permission or only the artificial intelligence terminal that satisfies the set condition has write permission, thereby avoiding behavior
  • the control data is maliciously modified by the artificial intelligence terminal or the unauthenticated artificial intelligence terminal, and the data can be updated.
  • the first behavior is evaluated according to the artificial intelligence ethics system, the first executable degree is obtained, and the first behavior and the first executable degree are stored in the behavior control database, because the behavior control database
  • the data is evaluated by the artificial intelligence ethics system, so it can ensure that the data in the behavior control database conforms to the specifications and meets the user's behavior requirements for artificial intelligence. Therefore, the artificial intelligence terminal can use the data in the behavior control database to judge whether the behavior generated by the artificial intelligence terminal is feasible, and can use the data in the behavior control database to standardize the behavior of the terminal, so that the operation or action of the terminal is more reasonable and satisfied. User requirements for the terminal.
  • the processor 702 can utilize the deep learning neural network to evaluate the first behavior. As shown in FIG. 9, before executing 801, the processor 702 is further configured to perform:
  • the deep learning neural network is composed of a number of neurons, wherein each neuron includes at least one input and output and computational functions, each input having a set weight. Moreover, each input represents a behavioral feature, and the behavior characteristics of different inputs can be weighted, and the calculation function is input to perform calculation, and the executable degree of the behavior having the behavior characteristic corresponding to the input is output.
  • the neural network algorithm may include: a perceptron neural network (Perceptron) Neural Network), Back Propagation, Hopfield Network, Self-Organizing Map, SOM). Learning Vector Quantization (Learning Vector Quantization, LVQ). Further, the algorithm adopted by the neural network is a Restricted Boltzmann Machine (RBN), Deep Belief.
  • the deep learning neural network may not be created according to the artificial intelligence ethics system based on the principle of altruism, but may be constructed according to other setting standards according to actual needs.
  • the foregoing 803 may specifically include: using the deep learning neural network to evaluate each behavior characteristic of the first behavior, thereby obtaining a first executable degree of the first behavior.
  • each behavioral feature of the first behavior is input to a corresponding neuron in the neural network, and the neuron can weight each behavioral feature according to its input weight, and input a calculation function to calculate, and output a behavior with a corresponding input.
  • the enforceability of the behavior of the feature is a simple action, a simple action, a simple action, a simple action, a simple action, and a simple action.
  • the processor 102 can directly evaluate the behavior of the behavior according to the artificial intelligence ethics system and act as the behavior control data, and can directly receive the external behavior and the executable degree as the behavior control data.
  • the processor 702 is further configured to execute:
  • the processor 702 can obtain the behavior data of the second behavior of the sample input by the user and the second executable degree of the second behavior by the input/output device 701 such as the machine interaction interface, the remote control device, the microphone, or the like; or the processor 702
  • the behavior data of the second behavior as a sample sent by the other device and the second executable degree of the second behavior may be received by the receiver.
  • the other device may be an artificial intelligence terminal or a server for managing the artificial intelligence terminal.
  • the behavior processing device 700 may be the aforementioned artificial intelligence terminal or server, and the behavior processing device 700 may also be used to execute the previous behavior control method after performing the foregoing establishment method.
  • an embodiment of the present invention further provides a computer readable storage medium storing a program, the program being executable to implement a behavior control method or a method for establishing a behavior control database as described above.
  • the program may be stored in a computer readable storage medium, and the storage medium may include: Flash drive, read-only memory (Read-Only Memory, ROM), Random Access Memory (RAM), disk or CD.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Robotics (AREA)
  • Mechanical Engineering (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Algebra (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Fuzzy Systems (AREA)
  • Evolutionary Computation (AREA)
  • Biomedical Technology (AREA)
  • Mathematical Analysis (AREA)
  • Molecular Biology (AREA)
  • Pure & Applied Mathematics (AREA)
  • Computing Systems (AREA)
  • Automation & Control Theory (AREA)
  • Mathematical Physics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Manipulator (AREA)

Abstract

本发明实施例涉及计算机技术领域，公开了一种人工智能终端系统、服务器及其行为控制方法。其中，该方法包括：接收人工智能终端上传的待执行的第一行为数据；将待执行的第一行为数据与已经建立的行为控制数据库中的行为控制数据进行匹配，得到第一可执行度；将所述第一可执行度发送至所述人工智能终端，以使所述人工智能终端根据所述第一可执行度判断人工智能终端是否可执行所述第一行为数据。本发明实施例，可以利用行为控制数据对人工智能终端的行为进行规范。

Description

人工智能终端系统、服务器及其行为控制方法
【技术领域】
本发明涉及计算机技术领域，具体涉及一种人工智能终端系统、服务器及其行为控制方法。
【背景技术】
人工智能（AI，Artificial Intelligence）是研究、开发用于模拟、延伸和扩展人的智能的理论、方法、技术及应用系统的一门新的技术科学。人工智能是计算机科学的一个分支，它企图了解智能的实质，并生产出一种新的能以人类智能相似的方式做出反应的智能机器，该领域的研究包括机器人、语言识别、图像识别、自然语言处理和专家系统等。从1956年正式提出人工智能学科算起，50多年来，取得长足的发展，成为一门广泛的交叉和前沿科学。时至今日，人工智能的发展已经渗透到社会生活中的很多层面，把人类从繁重的体力中解放出来，同时也在逐步解放人类的脑力劳动。
本发明的发明人发现，现有的智能终端，目前还处于弱人工智能阶段，多数还是依赖于用户的控制而执行相应的行为，少数的智能终端能够针对外界指令结合环境因素自行执行一些行为，但这些行为有时不太合理，并不能满足用户对终端的要求。
【发明内容】
本发明实施例所要解决的技术问题是提供一种人工智能终端系统、服务器及其行为控制方法，用于利用行为控制数据对人工智能终端的行为进行规范。
本发明实施例提供一种人工智能终端的行为控制方法，所述方法包括：接收人工智能终端上传的待执行的第一行为的行为数据；将待执行的第一行为的行为数据与已经建立的行为控制数据库中的行为控制数据进行匹配，得到第一匹配结果；将所述第一匹配结果发送至所述人工智能终端，以使所述人工智能终端根据所述第一匹配结果判断人工智能终端是否可执行所述第一行为。
相应的，本发明实施例还提供一种服务器，包括输入装置、输出装置、存储器和处理器，所述存储器中存储有行为控制数据，所述处理器用于执行：利用所述输入装置接收人工智能终端上传的第一行为的行为数据；将待执行的第一行为的行为数据与已经建立的行为控制数据库中的行为控制数据进行匹配，得到第一匹配结果；利用所述输出装置将所述第一匹配结果发送至所述人工智能终端，以使所述人工智能终端根据所述第一匹配结果判断人工智能终端是否可执行所述第一行为。
相应的，本发明实施例还提供一种人工智能控制系统，所述系统包括如前所述的服务器及与所述服务器通讯连接的人工智能终端。
相应的，本发明实施例还提供一种计算机可读存储介质，所述计算机存储介质存储有程序，所述程序能够被执行以实现如前所述的行为控制方法。
本发明实施例中，通过预先在与人工智能终端连接的服务器中存储行为控制数据，并利用行为控制数据判断人工智能终端的行为是否可行，可以利用行为控制数据对终端的行为进行规范，使终端的操作或动作更为合理，满足用户对终端的要求。
【附图说明】
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是本发明实施例中人工智能终端的结构示意框图；
图2是本发明实施例中人工智能控制系统的结构示意框图；
图3是本发明实施例中行为控制方法的第一流程示意图；
图4是本发明实施例中行为控制方法的第二流程示意图；
图5是本发明实施例中服务器的结构示意框图；
图6是本发明实施例中行为控制方法的第三流程示意图；
图7是本发明实施例中行为处理设备的结构示意框图；
图8是本发明实施例中行为控制数据库的建立方法的第一流程示意图；
图9是本发明实施例中行为控制数据库的建立方法的第二流程示意图;
图10是本发明实施例中行为控制数据库的建立方法的部分流程示意图。
【具体实施方式】
下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
本发明的说明书和权利要求书及上述附图中的术语“第一”、“第二”、“第三”“第四”等（如果存在）是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的本发明的实施例，例如能够以除了在这里图示或描述的那些以外的顺序实施。此外，术语“包括”和“具有”以及他们的任何变形，意图在于覆盖不排他的包含，例如，包含了一系列步骤或单元的过程、方法、系统、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。
本发明实施例提供一种人工智能终端及相应的行为控制方法，以下进行具体说明：
如图1所示，本发明实施例提供一种人工智能终端100，该智能终端可以包括输入输出装置101、处理器102和存储器103，输入输出装置101和存储器103分别与处理器102连接，具体可采用总线连接方式。其中存储器103中存储有行为控制数据。可以理解的是，人工智能终端100也可不设置输入输出装置101，在此不作限定。
存储器103中还可以存储有程序代码，处理器能够调用所述程序代码并执行如下方法：
S101、获取待执行的第一行为的行为数据；
例如，智能终端100还可以包括传感器，可用于从环境中实时采集数据；举例来说，传感器可以包括图像传感器、声音传感器、气味传感器、温度传感器、湿度传感器、距离传感器、位置传感器等一种或多种，从而采集各种环境数据；在采集到环境数据后，处理器可以对环境数据进行处理，从而生成待执行的第一行为的行为数据，以对周围环境做出应对；
又例如，接收用户输入的指令或者其他设备如服务器发送的指令，并根据该指令解析得到第一行为的行为数据。
S102、将待执行的第一行为的行为数据与预先存储的行为控制数据进行匹配，得到第一匹配结果；
其中，所述行为控制数据可以是预先依据以利他为原则的人工智能道德体系所建立的数据；
S103、根据所述第一匹配结果控制所述第一行为的执行。
需要说明的是，第一行为可以是智能终端待执行的操作、动作、计算、输入、输出等行为中的任意一种或多种的组合；
举例来说，行为控制数据可以是预先存储在存储器103中的数据，其中可包括多个行为和对应的可执行度，优选的，多个行为可以被分为若干类别，例如按照儒、释、道、医、武、乐、俗、科、商……分别分类进行存储。
在一实施例中，该第一匹配结果可以明确为匹配（可执行）或不匹配（不可执行）两种结果。例如，处理器102判断在预先存储的行为控制数据中是否存在与第一行为的行为数据匹配的可执行行为控制数据；如存在则结果为匹配，则控制执行第一行为；否则为不匹配，则控制不执行该第一行为。
在另一实施例，该第一匹配结果可包括可执行度。例如，处理器102将第一行为的行为数据与预先存储的行为控制数据进行匹配，得到第一行为的第一可执行度；并根据所述第一可执行度判断人工智能终端是否可执行所述第一行为。进而，处理器102可根据判断结果控制人工智能终端是否执行该第一行为。
例如，当智能终端将要执行某一行为时，处理器102将利用行为控制数据对该行为的行为数据进行匹配，并根据匹配得到的可执行度来判断该行为是否可执行，具体的判断方法可以是：预先设置一阈值，若可执行度高于该阈值的行为，则判断为可执行，反之，则判断为不可执行。该阈值的具体数值可以根据经验值进行设置，并且可以动态调整。
本实施例中提供的智能终端，通过预先在终端中存储行为控制数据，可以利用行为控制数据对终端的操作或动作进行规范，使终端的操作或动作更为合理，满足用户对终端的要求。
举例来说，智能终端100可以是手机、电脑、智能机器人或其他具备数据分析处理能力的终端，下面以智能机器人为例进行说明：
智能机器人的存储器中存储有行为控制数据，在一种实施方式中，该行为控制数据是固化在机器人存储器中的只读数据，可防止该行为控制数据被恶意修改；在另外一种实施方式中，智能机器人可以具备自学习功能，并利用自学习自行地对行为控制数据进行更新和修正。
智能机器人根据用途不同，可以有多种分类，具备各不相同的功能，例如家用机器人可具备移动、抓取/搬运物品、语音对话等功能，工业机器人又可根据不同的用途配置相应的功能，例如焊接、搬运、喷漆、打磨、码垛、装配等，军用机器人可具备复杂地形移动、负重、排雷、射击等功能……。
以上种种智能机器人可执行多种人类难以胜任的繁重工作，但机器人在工作过程中，时常要面对一些复杂或突发情况，对于复杂或突发情况可以至少设计如下两种处理方式：一种处理方式是机器人停止工作并需求人类用户的指示，这种处理方式较为稳妥，但显然很可能会造成机器人工作的延误，浪费机器人的处理能力；第二种处理方式是赋予机器人一定的自主处理权限，可对其面临的复杂或突发情况进行反应，执行合适的操作或动作对情况进行处理，但会带来一定的风险：机器人对复杂或突发情况的处理可能与人类用户的预期并不一致，甚至可能造成极为负面的后果。
本发明实施例提供的方案中，可在智能机器人的存储器中预先存储行为控制数据，机器人在面对复杂或突发情况时，其生成的处理方案（也即待执行的行为）的相关数据需要经过与行为控制数据进行匹配，只有匹配的方案才能够被执行，或进一步为只有可执行度高的方案才能够被执行，而行为控制数据是人类用户定义的，其中可对人类认为合理的行为赋予高可执行度，对于不合理的行为赋予低可执行度，只要行为控制数据的内容合理，则可控制具有一定自主处理能力的智能机器人在面对复杂或突发情况时的处理策略，防止机器人不合理应对造成的负面后果。
如图2所示，本发明实施例还提供一种人工智能控制系统，该系统包括前文所述的人工智能终端100以及服务器200，该服务器200以有线或无线的方式与人工智能终端100通讯连接，例如通过互联网连接，服务器200可以部署在云端，其中存储有行为控制数据库，行为控制数据库包括海量的行为控制数据，并且可以不断更新。
优选的，该输入输出装置101可以用于将终端待执行的第一行为的行为数据上传至服务器200，服务器200可将第一行为的行为数据与行为控制数据库进行匹配，得到该行为的第二可执行度，并将第二可执行度反馈给终端；终端在接收到第二可执行度后，可利用第一可执行度与第二可执行度进行第一行为是否可执行的判断，例如，将第一可执行度与第二可执行度进行加权平均后得到第三可执行度，再判断第三可执行度是否大于预定的阈值，如果大于则判断第一行为可执行，否则判断为不可执行。
机器人的存储器中存储的行为控制数据可能数据量较为有限，则其对行为的匹配结果可能不够准确，本发明实施例还提供一种方案，使得机器人可将其生成的处理方案上传给服务器，再接收服务器反馈的可执行度，以做更准确的判断。由于服务器中可存储有海量的行为控制数据，并且是不断补充和修正的，服务器对行为的判断可以认为是更可靠和准确的，这样借助于服务器反馈的可执行度，机器人可更准确的判断其生成的处理方案是否可执行。鉴于此，机器人在对第一可执行度和服务器反馈的第二可执行度进行加权平均时，可以对第二可执行度赋予更高的权重。
如图3所示，本发明实施例还提供一种行为控制方法，该方法可以应用于前述的人工智能终端100，具体可由上述人工智能终端100的处理器102执行。该方法可以包括：
301、获取待执行的第一行为的行为数据；
需要说明的是，第一行为可以是智能终端100待执行的任意操作或动作；
302、将待执行的第一行为的行为数据与预先存储的行为控制数据进行匹配，得到第一匹配结果；
举例来说，行为控制数据可以是预先存储在存储器中的数据，其中可包括多个行为，优选的，多个行为可以被分为若干类别，分别进行存储；进一步地，行为控制数据还可包括其包含的各行为对应的可执行度。
303、根据所述第一匹配结果控制所述第一行为的执行。
例如，若该第一匹配结果包括第一行为的第一可执行度，则根据所述第一可执行度判断是否可执行所述第一行为。当智能终端将要执行某一行为时，将利用行为控制数据对该行为的的行为数据进行匹配，并根据匹配得到的可执行度来判断该行为是否可执行，具体的判断方法可以是：预先设置一阈值，若可执行度高于该阈值的行为则判断为可执行，反之，则判断为不可执行。该阈值的具体数值可以根据经验值进行设置，并且可以动态调整。
本实施例中提供的行为处理方法，通过预先存储的行为控制数据来进行行为可执行度的匹配，可以利用行为控制数据对终端的操作或动作进行规范，使终端的操作或动作更为合理，满足用户对终端的要求。
举例来说，人工智能终端可以利用特征提取和特征匹配的方式来获取第一行为的第一可执行度，具体而言，上述S302可包括：
302A、提取所述第一行为的行为特征数据；
具体如，对所述第一行为的行为数据进行解析，以从该行为数据中提取得到第一行为的行为特征数据。
302B、将所述第一行为的行为特征数据与所述行为控制数据进行特征匹配，得到第一匹配结果。
其中，所述行为特征可包括下列特征中的至少一种：行为时间、行为地点、行为主体、行为对象。
上述302B具体可包括在行为控制数据中查找包含所述第一行为的各行为特征的至少一个行为的行为数据，根据所述至少一个行为的可执行度获得所述第一行为的第一可执行度。进一步地，所述根据所述至少一个行为的可执行度获得所述第一行为的第一可执行度包括：如果所述至少一个行为为一个行为，则将该行为的可执行度作为所述第一行为的第一可执行度；如果所述至少一个行为包括多个行为，则将多个行为的多个可执行度进行加权平均，得到所述第一行为的第一可执行度。
优选的，如图4所示，该方法还可包括：
401、将所述第一行为的行为数据上传至服务器，以使所述服务器将所述第一行为的行为数据与服务器中存储的行为控制数据库进行匹配获得第二匹配结果，并接收所述服务器反馈的第二匹配结果；
该服务器可以部署在云端，其中存储有行为控制数据库，行为控制数据库包括海量的行为控制数据，并且可以不断更新。且该服务器中的行为控制数据库同理于人工智能终端的行为控制数据库，为预先依据以利他为原则的人工智能道德体系所建立的数据。服务器中的行为控制数据库中的行为种类可设置为比人工智能终端的行为控制数据中的行为种类丰富。而且，服务器获得第二匹配度的方式类似于上述获得第一匹配结果的方式，同理于上述第一匹配结果，该第二匹配结果可以明确为匹配（可执行）或不匹配（不可执行）两种结果，或者可包括第二可执行度，在此不作赘述。
该服务器中的行为控制数据可为预先存储在执行所述方法的服务器中的只读数据；或者为预先存储在执行所述方法的服务器中的可读可写数据，但所有所述人工智能终端均不具有写入权限；又或者为预先存储在执行所述方法的服务器中的可读可写数据，但仅满足设定条件的人工智能终端具有写入权限。
则前述303可包括：根据所述第一匹配结果和所述第二匹配结果控制所述第一行为的执行。例如为，根据所述第一可执行度及第二可执行度判断是否可执行所述第一行为。
具体地，前述303可以包括：
303A、将所述第一可执行度与所述第二可执行度进行加权平均获得第三可执行度；
303B、判断所述第三可执行度是否大于预定的阈值，如果是，则执行301C否则执行303D；
303C、可执行所述第一行为；结束流程；
303D、不可执行所述第一行为。
举例来说，前述301可以包括：
301A1、利用传感器从环境中采集数据；
301B1、对所述传感器采集的数据进行处理，生成待执行的所述第一行为的行为数据。
智能终端100还可以包括传感器，可用于从环境中实时采集数据；举例来说，传感器可以包括图像传感器、声音传感器、气味传感器、温度传感器、湿度传感器、距离传感器、位置传感器等一种或多种，从而采集各种环境数据；在采集到环境数据后，处理器可以对环境数据进行处理，从而生成待执行的第一行为的行为数据，以对周围环境做出应对。
另外，本发明实施例还提供一种计算机可读存储介质，该计算机存储介质存储有程序，所述程序能够被执行以实现如前所述的行为处理方法。
举例来说，执行上述行为处理方法的智能终端100可以是手机、电脑、智能机器人或其他具备数据分析处理能力的终端，下面以智能机器人为例进行说明：
智能机器人的存储器中存储有行为控制数据，优选的，该行为控制数据是固化在机器人存储器中的只读数据，可防止该行为控制数据被恶意修改。
智能机器人根据用途不同，可以有多种分类，具备各不相同的功能，例如家用机器人可具备移动、抓取/搬运物品、语音对话等功能，工业机器人又可根据不同的用途配置相应的功能，例如焊接、搬运、喷漆、打磨、码垛、装配等，军用机器人可具备复杂地形移动、负重、排雷、射击等功能……。
以上种种智能机器人可执行多种人类难以胜任的繁重工作，但机器人在工作过程中，时常要面对一些复杂或突发情况，对于复杂或突发情况可以至少设计如下两种处理方式：一种处理方式是机器人停止工作并需求人类用户的指示，这种处理方式较为稳妥，但显然很可能会造成机器人工作的延误，浪费机器人的处理能力；第二种处理方式是赋予机器人一定的自主处理权限，可对其面临的复杂或突发情况进行反应，执行合适的操作或动作对情况进行处理，但会带来一定的风险：机器人对复杂或突发情况的处理可能与人类用户的预期并不一致，甚至可能造成极为负面的后果。
本发明实施例提供的处理方法中，智能机器人预先存储行为控制数据，机器人在面对复杂或突发情况时，其生成的处理方案（也即待执行的行为）的相关数据需要经过与行为控制数据进行匹配，只有匹配的方案才能够被执行，或进一步为只有可执行度高的方案才能够被执行，而行为控制数据是人类用户定义的，其中可对人类认为合理的行为赋予高可执行度，对于不合理的行为赋予低可执行度，只要行为控制数据的内容合理，则可控制具有一定自主处理能力的智能机器人在面对复杂或突发情况时的处理策略，防止机器人不合理应对造成的负面后果。
机器人的存储器中存储的行为控制数据可能数据量较为有限，则其对行为的匹配结果可能不够准确，本发明实施例还提供一种方案，使得机器人可将其生成的处理方案上传给服务器，再接收服务器反馈的可执行度，以做更准确的判断。由于服务器中可存储有海量的行为控制数据，并且是不断补充和修正的，服务器对行为的判断可以认为是更可靠和准确的，这样借助于服务器反馈的可执行度，机器人可更准确的判断其生成的处理方案是否可执行。鉴于此，机器人在对第一可执行度和服务器反馈的第二可执行度进行加权平均时，可以对第二可执行度赋予更高的权重。
如图5所示，本发明实施例还提供一种服务器200，该服务器200如前述的服务器，连接于前述人工智能终端，可用于管理该人工智能终端。具体，服务器200包括输入输出装置201、处理器202和存储器203。输入输出装置201和存储器203分别与处理器202连接，具体可采用总线连接方式。该输入输出装置101可包括用于接收器、发送器以及触摸屏、按键、数据接口如USB接口、显示屏等可与外界进行信息交互的装置。其中存储器203中存储有行为控制数据。
存储器203中还可以存储有程序代码，如图6所示，处理器202能够调用所述程序代码并执行如下方法：
601、接收人工智能终端上传的待执行的第一行为的行为数据。
具体，处理器104通过输入输出装置101接收到人工智能终端上传的第一行为的行为数据。其中，该人工智能终端可如前面实施例所述获取该第一行为的行为数据。例如，在601之前，处理器202通过输入输出装置601向人工智能终端发送行为指令，以使人工智能终端解析所述行为指令，得到所述第一行为的行为数据。
602、将待执行的第一行为的行为数据与已经建立的行为控制数据库中的行为控制数据进行匹配，得到第一匹配结果。
其中，所述行为控制数据库可为预先依据以利他为原则的人工智能道德体系所建立的数据库。举例来说，行为控制数据可以是预先存储在存储器104中的数据，其中可包括多个行为，还可包括对应的可执行度，优选的，多个行为可以被分为若干类别，例如按照儒、释、道、医、武、乐、俗、科、商……分别分类进行存储。
在一实施例中，该第一匹配结果可以明确为匹配（可执行）或不匹配（不可执行）两种结果。例如，处理器202判断在预先存储的行为控制数据中是否存在与第一行为的行为数据匹配的可执行行为控制数据；其中，如存在则结果为匹配，则表示可执行第一行为；否则为不匹配，则表示不可执行该第一行为。
在另一实施例，该第一匹配结果可包括可执行度。例如，处理器202将第一行为的行为数据与预先存储的行为控制数据进行匹配，得到第一行为的第一可执行度。
603、将所述第一匹配结果发送至所述人工智能终端，以使所述人工智能终端根据所述第一匹配结果判断人工智能终端是否可执行所述第一行为。
具体，人工智能终端可如前面实施例所述判断是执行该第一行为。
本实施例中提供的用于管理人工智能终端的服务器，通过预先在服务器中存储行为控制数据，可以利用行为控制数据对终端的操作或动作进行规范，使终端的操作或动作更为合理，满足用户对终端的要求。
优选的，该行为控制数据可以是存储器203中存储的只读数据，可以防止该行为控制数据被恶意修改。或者，该行为控制数据为预先存储在存储器203中的可读可写数据，但与该服务器连接的所有人工智能终端均不具有写入权限，而服务器仍可具有修改权限，由此既可防止该行为控制数据被人工智能终端恶意修改，也可保证数据的可更新。又或者，该行为控制数据为预先存储在执行存储器203中的可读可写数据，但仅满足设定条件的人工智能终端具有写入权限，而服务器和经认证的部分人工智能终端仍可具有修改权限，由此既可避免行为控制数据被未经认证的人工智能终端恶意修改，也可保证数据的可更新。其中，该设定条件可根据实际需求设定，例如为终端用户身份已通过验证，终端级别为安全级别等。
举例来说，服务器的存储器203中存储有行为控制数据，在一种实施方式中，该行为控制数据是固化在服务器存储器203中的只读数据，可防止该行为控制数据被恶意修改；在另外一种实施方式中，服务器可以具备自学习功能，并利用自学习功能自行地对行为控制数据进行更新和修正。
举例来说，服务器200可以利用特征提取和特征匹配的方式来获取第一匹配结果（如第一行为的第一可执行度），具体而言，处理器202执行的上述602可包括：
602A、对所述第一行为进行解析，提取所述第一行为的行为特征。
602B、将所述第一行为的行为特征与所述行为控制数据库中的行为控制数据进行特征匹配，从而获得所述第一匹配结果。
具体地，该行为控制数据库可存储有不同行为及其行为特征、不同行为的可执行度。该行为控制数据库中存储的行为的可执行度可由处理器503根据用户对终端历史行为的反馈进行自学习得到，或者由用户输入得到的。
当所述第一匹配结果包括第一可执行度时；处理器503执行的602B可包括：
602B1、在所述行为控制数据库中查找包含所述第一行为的各行为特征的至少一个第二行为的行为数据。
602B2、根据所述至少一个第二行为的可执行度获得所述第一行为的第一可执行度。
例如，该602B2具体可为：如果所述至少一个第二行为仅有一个第二行为，则将该第二行为的可执行度直接作为得到所述第一行为的第一可执行度；如果所述至少一个第二行为包括多个第二行为，则将多个第二行为的多个可执行度进行加权平均，得到所述第一行为的第一可执行度。
其中，所述行为特征可包括下列特征中的至少一种：行为时间、行为地点、行为主体、行为对象。
优选的，人工智能终端可如前实施例所述先将待执行的第一行为的行为数据与自身存储的行为控制数据进行匹配，得到预判执行度，其中其预判执行度的获得过程类似于服务器的上述第一执行度的获取；人工智能终端在其预判执行度小于第一预定阈值时，直接判断该第一行为不可执行，否则将终端待执行的第一行为的行为数据上传至服务器200；服务器200将第一行为的行为数据与其存储的行为控制数据库进行匹配，得到该行为的实际可执行度，并将实际可执行度反馈给终端；终端在接收到实际可执行度后，判断实际可执行度是否大于第二预定阈值，如果大于则判断第一行为可执行，否则判断为不可执行。其中，该第一预定阈值和第二预定阈值可根据实际情况设置，其数值关系可设置为第一预定阈值小于第二预定阈值。
由于将人工智能终端的存储器中存储的行为控制数据可能数据量较为有限，则其对行为的匹配结果可能不够准确，故可将终端经出判断后的第一行为的行为数据上传给服务器，再接收服务器反馈的可执行度，以做更准确的判断。由于服务器中可存储有海量的行为控制数据，并且是不断补充和修正的，服务器对行为的判断可以认为是更可靠和准确的，这样借助于服务器反馈的可执行度，机器人可更准确的判断其生成的处理方案是否可执行。鉴于此，机器人在对第一可执行度和服务器反馈的第二可执行度进行加权平均时，可以对第二可执行度赋予更高的权重。
本发明实施例还提供一种行为处理设备及相应的建立方法，以下进行具体说明：
如图7所示的是一种行为处理设备700，行为处理设备700可以包括输入输出装置701、处理器702和存储器703。其中，输入输出装置701、存储器703均与处理器702连接，具体可采用总线连接方式。存储器703中存储有行为控制数据库，该行为控制数据库用于存储前述行为控制数据，以提供给人工智能终端或连接于人工智能终端的服务器，以获取人工智能终端的待执行行为的可执行度，进而使得人工智能终端根据获取的可执行度判断是否执行该待执行行为。
存储器703中还可以存储有程序代码，如图8所示，处理器702能够调用所述程序代码并执行如下方法：
801、获取第一行为的行为数据。
具体，处理器703利用输入输出装置701获得第一行为的行为数据。其中，输入输出装置701为人机交互界面、遥控设备、麦克风、接收器、采集器等可从外界获得数据的任意装置，用于从外界获取行为数据。例如，处理器703直接获取用户通过人机交互界面、遥控设备、麦克风输入的行为数据，或者接收其他设备如其他人工智能终端或者服务器发送的行为数据，又或者获取如图像采集器等采集得到的行为数据。
举例来说，第一行为可以是人工智能终端可执行的操作、动作、计算、输入、输出等行为中的任意一种或多种的组合。
802、对所述第一行为进行解析，提取所述第一行为的行为特征。
其中，所述行为特征可包括下列特征中的至少一种：行为时间、行为地点、行为主体、行为对象。
803、对所述第一行为的各行为特征进行评价，从而获得所述第一行为的第一可执行度。
例如，可以按照以利他为原则的人工智能道德体系对所述第一行为的各行为特征进行评价，从而获得所述第一行为的第一可执行度。当然，也可按照其他设定评价原则继续该评价。
其中，该人工智能道德体系可以是以“社会主义荣辱观”、《弟子规》、《论语》、《道德经》等里面的利他行为准则为数据建立的体系。
具体，人工智能道德体系可以包含有儒、释、道、医、武、乐、俗、科、商……等不同类别的道德内容。处理器702将第一行为的行为特征与人工智能道德体系中的道德内容进行匹配，由其匹配度得到第一行为的可执行度。当第一行为的行为特征为多个，可预先为不同行为特征设置权重，并将每个行为特征的匹配度进行加权平均得到该第一行为的可执行度。
804、将所述第一行为以及相对应的所述第一可执行度存储在所述行为控制数据库中。
举例来说，行为控制数据库中存储的行为可以被分为若干类别，例如可先将该第一行为按照儒、释、道、医、武、乐、俗、科、商……分别分类后，再进行存储。该第一行为的分类可直接在执行803的可执行度评价过程中进行。
另外，所述行为控制数据库中的数据可被存储为只读数据，以避免后续使用该行为控制数据的设备（如人工智能终端或服务器）恶意修改。或者，该行为控制数据库中的数据被存储为可读可写数据，但所有人工智能终端均不具有写入权限或者仅满足设定条件的人工智能终端具有写入权限，由此既可避免行为控制数据被人工智能终端或未经认证的人工智能终端恶意修改，也可保证数据的可更新。
本发明实施例中，按照人工智能道德体系对第一行为进行评价，得到其第一可执行度，并将第一行为和第一可执行度存储为行为控制数据库中，由于行为控制数据库中的数据经过人工智能道德体系的评价，故可保证该行为控制数据库中的数据符合规范，且满足用户对人工智能的行为要求。由此，人工智能终端可利用行为控制数据库中的数据判断人工智能终端生成的行为是否可行，可以利用行为控制数据库中的数据对终端的行为进行规范，使终端的操作或动作更为合理，满足用户对终端的要求。
举例来说，处理器702可利用深度学习神经网络对第一行为进行评价。如图9所示，处理器702在执行上述801之前，还用于执行：
805、构建依据以利他为原则的人工智能道德体系的深度学习神经网络。
例如，该深度学习神经网络有若干个神经元组成，其中，每个神经元包括至少一个输入和输出以及计算功能，每个输入具有设定权值。且，每个输入表示一个行为特征，可将不同输入的行为特征进行加权后，并输入计算功能进行计算，输出具有对应输入的行为特征的行为的可执行度。具体，该神经网络算法可包括：感知器神经网络（Perceptron Neural Network）, 反向传递（Back Propagation）， Hopfield网络，自组织映射（Self-Organizing Map, SOM）。学习矢量量化（Learning Vector Quantization， LVQ）。进一步地，该神经网络采用的算法为受限波尔兹曼机（Restricted Boltzmann Machine， RBN），Deep Belief Networks（DBN），卷积网络（Convolutional Network）,堆栈式自动编码器（Stacked Auto-encoders）等深度学习算法，以在每次对第一行为评价的同时进行深度学习，如可通过学习来调整神经元的输入的权值等。当然，在其他实施例中，深度学习神经网络也可不依据以利他为原则的人工智能道德体系创建，而是依据根据实际需求根据其他设定标准进行构建。
并且，上述803具体可包括：利用所述深度学习神经网络对所述第一行为的各行为特征进行评价，从而获得所述第一行为的第一可执行度。
例如，将第一行为的各行为特征输入至神经网络中的相应神经元，该神经元可将各行为特征进行按照其输入权重进行加权后，并输入计算功能进行计算，输出具有对应输入的行为特征的行为的可执行度。
另外，处理器102除可根据人工智能道德体系自行对行为进行可执行度评价并作为行为控制数据外，还可直接接收外界的行为及其可执行度作为行为控制数据。如图10所示，处理器702还用于执行：
806、接收作为样本的第二行为的行为数据及其第二可执行度。
具体，处理器702可通过机交互界面、遥控设备、麦克风等输入输出装置701获取用户输入的作为样本的第二行为的行为数据及该第二行为的第二可执行度；或者，处理器702可通过接收器接收其他设备发送的作为样本的第二行为的行为数据及该第二行为的第二可执行度。该其他设备可以为人工智能终端或用于管理人工智能终端的服务器。
807、将所述第二行为的行为数据以及相对应的第二可执行度存储在所述行为控制数据库中。
举例而言，行为处理设备700可以为前述人工智能终端或服务器，该行为处理设备700在执行前述建立方法之后，还可以用于执行前面行为控制方法。
另外，本发明实施例还提供一种计算机可读存储介质，该计算机存储介质存储有程序，所述程序能够被执行以实现如前所述的行为控制方法或行为控制数据库的建立方法。
需要说明的是，以上各实施例均属于同一发明构思，各实施例的描述各有侧重，在个别实施例中描述未详尽之处，可参考其他实施例中的描述。
本领域普通技术人员可以理解上述实施例的各种方法中的全部或部分步骤是可以通过程序来指令相关的硬件来完成，该程序可以存储于一计算机可读存储介质中，存储介质可以包括：闪存盘、只读存储器（Read-Only Memory ，ROM）、随机存取器（Random Access Memory，RAM）、磁盘或光盘等。
以上对本发明实施例所提供的智能终端、行为处理方法及系统进行了详细介绍，本文中应用了具体个例对本发明的原理及实施方式进行了阐述，以上实施例的说明只是用于帮助理解本发明的方法及其核心思想；同时，对于本领域的一般技术人员，依据本发明的思想，在具体实施方式及应用范围上均会有改变之处，综上所述，本说明书内容不应理解为对本发明的限制。

Claims (21)

  1. 一种人工智能终端的行为控制方法，其特征在于，所述方法包括：
    接收人工智能终端上传的待执行的第一行为的行为数据；
    将待执行的第一行为的行为数据与已经建立的行为控制数据库中的行为控制数据进行匹配，得到第一匹配结果；
    将所述第一匹配结果发送至所述人工智能终端，以使所述人工智能终端根据所述第一匹配结果判断人工智能终端是否可执行所述第一行为。
  2. 根据权利要求1所述的方法，其特征在于，所述行为控制数据为预先依据以利他为原则的人工智能道德体系所建立的数据库。
  3. 根据权利要求1所述的方法，其特征在于，所述第一匹配结果包括所述第一行为的第一可执行度。
  4. 根据权利要求1至3任一项所述的方法，其特征在于，所述将待执行的第一行为的行为数据与已经建立的行为控制数据库中的行为控制数据进行匹配，得到第一匹配结果包括：
    对所述第一行为的行为数据进行解析，提取所述第一行为的行为特征；
    将所述第一行为的行为特征与所述行为控制数据库中的行为控制数据进行特征匹配，从而获得所述第一匹配结果。
  5. 根据权利要求4所述的方法，其特征在于，当所述第一匹配结果包括第一可执行度时；
    所述将所述第一行为的行为特征与所述行为控制数据库中的行为控制数据进行特征匹配，从而获得所述第一匹配结果包括：
    在所述行为控制数据库中查找包含所述第一行为的各行为特征的至少一个第二行为的行为数据，根据所述至少一个第二行为的可执行度获得所述第一行为的第一可执行度。
  6. 根据权利要求5所述的方法，其特征在于，所述根据所述至少一个第二行为的可执行度获得所述第一行为的第一可执行度包括：
    如果所述至少一个第二行为包括多个第二行为，则将多个第二行为的多个可执行度进行加权平均，得到所述第一行为的第一可执行度。
  7. 根据权利要求4所述的方法，其特征在于，所述行为特征包括下列特征中的至少一种：
    行为时间、行为地点、行为主体、行为对象。
  8. 根据权利要求1所述的方法，其特征在于，所述行为控制数据为预先存储在执行所述方法的服务器中的只读数据；或者
    所述行为控制数据为预先存储在执行所述方法的服务器中的可读可写数据，但所有所述人工智能终端均不具有写入权限；或者
    所述行为控制数据为预先存储在执行所述方法的服务器中的可读可写数据，但仅满足设定条件的人工智能终端具有写入权限。
  9. 根据权利要求1所述的方法，其特征在于，在所述接收人工智能终端上传的第一行为的行为数据之前，还包括：
    向所述人工智能终端发送行为指令，以使所述人工智能终端解析所述行为指令，得到所述第一行为的行为数据。
  10. 一种服务器，其特征在于，所述服务器包括输入装置、输出装置、存储器和处理器，所述存储器中存储有行为控制数据，所述处理器用于执行：
    利用所述输入装置接收人工智能终端上传的第一行为的行为数据；
    将待执行的第一行为的行为数据与已经建立的行为控制数据库中的行为控制数据进行匹配，得到第一匹配结果；
    利用所述输出装置将所述第一匹配结果发送至所述人工智能终端，以使所述人工智能终端根据所述第一匹配结果判断人工智能终端是否可执行所述第一行为。
  11. 根据权利要求10所述的服务器，其特征在于，所述行为控制数据库为预先依据以利他为原则的人工智能道德体系所建立的数据库。
  12. 根据权利要求10所述的服务器，其特征在于，所述第一匹配结果包括所述第一行为的第一可执行度。
  13. 根据权利要求10至12任一项所述的服务器，其特征在于，所述将待执行的第一行为的行为数据与已经建立的行为控制数据库中的行为控制数据进行匹配，得到第一匹配结果包括：
    对所述第一行为的行为数据进行解析，提取所述第一行为的行为特征；
    将所述第一行为的行为特征与所述行为控制数据库中的行为控制数据进行特征匹配，从而获得所述第一匹配结果。
  14. 根据权利要求13所述的服务器，其特征在于，当所述第一匹配结果包括第一可执行度时；
    所述将所述第一行为的行为特征与所述行为控制数据库中的行为控制数据进行特征匹配，从而获得所述第一匹配结果包括：
    在所述行为控制数据库中查找包含所述第一行为的各行为特征的至少一个第二行为的行为数据，根据所述至少一个第二行为的可执行度获得所述第一行为的第一可执行度。
  15. 根据权利要求14所述的服务器，其特征在于，所述根据所述至少一个第二行为的可执行度获得所述第一行为的第一可执行度包括：
    如果所述至少一个第二行为包括多个第二行为，则将多个第二行为的多个可执行度进行加权平均，得到所述第一行为的第一可执行度。
  16. 根据权利要求13所述的服务器，其特征在于，所述行为特征包括下列特征中的至少一种：
    行为时间、行为地点、行为主体、行为对象。
  17. 根据权利要求10所述的服务器，其特征在于，所述行为控制数据为预先存储在执行所述方法的服务器中的只读数据；或者
    所述行为控制数据为预先存储在执行所述方法的服务器中的可读可写数据，但所有所述人工智能终端均不具有写入权限；或者
    所述行为控制数据为预先存储在执行所述方法的服务器中的可读可写数据，但仅满足设定条件的人工智能终端具有写入权限。
  18. 根据权利要求10所述的服务器，其特征在于，所述处理器还用于：
    利用所述输出装置向所述人工智能终端发送行为指令，以使所述人工智能终端解析所述行为指令，得到所述第一行为的行为数据。
  19. 根据权利要求10所述的服务器，其特征在于，所述人工智能终端为智能机器人。
  20. 一种人工智能控制系统，所述系统包括如权利要求10-19任一项所述的服务器及与所述服务器通讯连接的人工智能终端。
  21. 一种计算机存储介质，其特征在于，所述计算机存储介质中存储有程序，所述程序能够被执行以实现如权利要求1-9任一项所述的人工智能终端的行为控制方法。
PCT/CN2017/099015 2017-08-25 2017-08-25 人工智能终端系统、服务器及其行为控制方法 WO2019037076A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2017/099015 WO2019037076A1 (zh) 2017-08-25 2017-08-25 人工智能终端系统、服务器及其行为控制方法
CN201780036358.1A CN109313645B (zh) 2017-08-25 2017-08-25 人工智能终端系统、服务器及其行为控制方法

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2017/099015 WO2019037076A1 (zh) 2017-08-25 2017-08-25 人工智能终端系统、服务器及其行为控制方法

Publications (1)

Publication Number Publication Date
WO2019037076A1 true WO2019037076A1 (zh) 2019-02-28

Family

ID=65225747

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/099015 WO2019037076A1 (zh) 2017-08-25 2017-08-25 人工智能终端系统、服务器及其行为控制方法

Country Status (2)

Country Link
CN (1) CN109313645B (zh)
WO (1) WO2019037076A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113741183A (zh) * 2021-08-12 2021-12-03 桂林电子科技大学 基于阻尼比模型的工业机器人自适应导纳控制方法

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111371842A (zh) * 2020-02-19 2020-07-03 上海智知盾科技有限公司 控制人工智能终端行为的方法和系统
CN116360598B (zh) * 2023-03-30 2024-03-26 无锡超体生命科技有限公司 基于脑电波的运动想象动作控制方法及系统

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030101151A1 (en) * 2001-11-26 2003-05-29 Holland Wilson Lee Universal artificial intelligence software program
CN101681446A (zh) * 2007-03-01 2010-03-24 波音公司 人类行为建模和仿真架构
CN103001947A (zh) * 2012-11-09 2013-03-27 北京奇虎科技有限公司 一种程序处理方法和系统
CN104361282A (zh) * 2014-10-31 2015-02-18 中国联合网络通信集团有限公司 移动终端安全防护方法及装置
US20150127593A1 (en) * 2013-11-06 2015-05-07 Forever Identity, Inc. Platform to Acquire and Represent Human Behavior and Physical Traits to Achieve Digital Eternity
CN104769645A (zh) * 2013-07-10 2015-07-08 哲睿有限公司 虚拟伴侣
CN205913705U (zh) * 2016-07-15 2017-02-01 克斯福佑株式会社 人工智能自动速度调节跑步机

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20150091569A (ko) * 2014-02-03 2015-08-12 삼성전자주식회사 전자 장치 및 이의 특정 영역에 대한 접근을 제어하는 방법
US11003748B2 (en) * 2015-12-28 2021-05-11 Unbotify Ltd. Utilizing behavioral features to identify bot
CN106383450A (zh) * 2016-11-10 2017-02-08 北京工商大学 一种基于大数据的智能家居用户行为分析系统及方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030101151A1 (en) * 2001-11-26 2003-05-29 Holland Wilson Lee Universal artificial intelligence software program
CN101681446A (zh) * 2007-03-01 2010-03-24 波音公司 人类行为建模和仿真架构
CN103001947A (zh) * 2012-11-09 2013-03-27 北京奇虎科技有限公司 一种程序处理方法和系统
CN104769645A (zh) * 2013-07-10 2015-07-08 哲睿有限公司 虚拟伴侣
US20150127593A1 (en) * 2013-11-06 2015-05-07 Forever Identity, Inc. Platform to Acquire and Represent Human Behavior and Physical Traits to Achieve Digital Eternity
CN104361282A (zh) * 2014-10-31 2015-02-18 中国联合网络通信集团有限公司 移动终端安全防护方法及装置
CN205913705U (zh) * 2016-07-15 2017-02-01 克斯福佑株式会社 人工智能自动速度调节跑步机

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113741183A (zh) * 2021-08-12 2021-12-03 桂林电子科技大学 基于阻尼比模型的工业机器人自适应导纳控制方法
CN113741183B (zh) * 2021-08-12 2023-10-27 桂林电子科技大学 基于阻尼比模型的工业机器人自适应导纳控制方法

Also Published As

Publication number Publication date
CN109313645B (zh) 2022-05-24
CN109313645A (zh) 2019-02-05

Similar Documents

Publication Publication Date Title
WO2019037076A1 (zh) 人工智能终端系统、服务器及其行为控制方法
WO2019037077A1 (zh) 人工智能的行为控制数据库的建立方法及其设备、系统
Yu et al. Adaptive NN impedance control for an SEA-driven robot
EP3413212A1 (en) Random forest model training method, electronic apparatus and storage medium
WO2019209059A1 (en) Machine learning on a blockchain
TW202209196A (zh) 用於連續小樣本學習的方法以及使用者設備
WO2019001558A1 (zh) 一种人机识别的方法和设备
Ma et al. Adaptive consensus of uncertain switched nonlinear multi-agent systems under sensor deception attacks
WO2019037075A1 (zh) 人工智能终端、系统及其行为控制方法
WO2019107674A1 (en) Computing apparatus and information input method of the computing apparatus
WO2020180008A1 (en) Method for processing plans having multiple end points and electronic device applying the same method
WO2019029061A1 (zh) 人工智能设备、系统及其行为控制方法
Wang et al. Review on the application of artificial intelligence in antivirus detection system i
RU2704538C1 (ru) Сетевая архитектура человекоподобной сети и способ реализации
Xu-Yang et al. Robust adaptive synchronization of chaotic neural networks by slide technique
WO2023033194A1 (ko) 가지치기 기반 심층 신경망 경량화에 특화된 지식 증류 방법 및 시스템
Zhang Application of Artificial Intelligence Technology in Computer Network Security.
Gavrylenko et al. Development of the disable software reporting system on the basis of the neural network
Ou Multiagent-based computer virus detection systems: abstraction from dendritic cell algorithm with danger theory
WO2020111704A1 (en) Electronic device for scheduling a plurality of tasks and operating method thereof
JP2021015421A (ja) 情報処理プログラム、情報処理方法および情報処理装置
KR20200094002A (ko) IoT 서비스 플랫폼 장치
RU2642406C1 (ru) Способ и система динамической идентификации личности испытуемого
WO2024053856A1 (ko) 메타 렌즈를 이용하여 획득한 이미지를 처리하는 전자 장치 및 그 동작 방법
Belikov et al. Calculation of the control signal in MIMO NN-based ANARX models: Analytical approach

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17922718

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 24.09.2020)

122 Ep: pct application non-entry in european phase

Ref document number: 17922718

Country of ref document: EP

Kind code of ref document: A1