WO2017166990A1 - 具备评价能力的人工智能系统及其评价方法 - Google Patents

具备评价能力的人工智能系统及其评价方法 Download PDF

Info

Publication number
WO2017166990A1
WO2017166990A1 PCT/CN2017/076267 CN2017076267W WO2017166990A1 WO 2017166990 A1 WO2017166990 A1 WO 2017166990A1 CN 2017076267 W CN2017076267 W CN 2017076267W WO 2017166990 A1 WO2017166990 A1 WO 2017166990A1
Authority
WO
WIPO (PCT)
Prior art keywords
evaluation
output
autonomous
negative
module
Prior art date
Application number
PCT/CN2017/076267
Other languages
English (en)
French (fr)
Inventor
刘若鹏
胡滨
Original Assignee
深圳光启合众科技有限公司
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 深圳光启合众科技有限公司 filed Critical 深圳光启合众科技有限公司
Publication of WO2017166990A1 publication Critical patent/WO2017166990A1/zh

Links

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
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • 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
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor

Definitions

  • the invention relates to the field of artificial intelligence, in particular to an artificial intelligence system with evaluation capability, and a method for evaluating the target object by the artificial intelligence system.
  • Artificial intelligence is a new technical science that studies theories, methods, techniques, and applications that simulate, extend, and extend human intelligence. Since its birth, the theory and technology of artificial intelligence has become more and more mature, and the field of application has been expanding. As a branch of computer science, it attempts to understand the essence of intelligence and produce a new kind of intelligence that can respond in a way similar to human intelligence. machine.
  • the invention provides an artificial intelligence system with evaluation ability and related evaluation method, and constructs an evaluation system of the artificial intelligence system for external things, so that the artificial intelligence system can have the self-evaluation target object with the increase of life experience. ability.
  • an artificial intelligence system with self-evaluation capability comprising: an evaluation capability module configured to generate, according to original evaluation information of a reference person for a predefined thing, Autonomous evaluation criteria for predefined things, the evaluation capability module is further configured to autonomously evaluate the target object according to the autonomous evaluation criteria; and a decision module configured to determine an output for the target object based on the autonomous evaluation.
  • an artificial intelligence system having self-evaluation capability comprising: an output module configured to perform output for a target object; and an evaluation capability module configured to pre-define according to a reference population The original evaluation information of the thing generates an autonomous evaluation criterion for the predefined thing, the evaluation capability module is further configured to perform an independent evaluation on the target object according to the self-evaluation criterion; and a decision module configured to adjust the output module according to the autonomous evaluation The output for the target thing.
  • an autonomous evaluation method for an artificial intelligence system comprising: generating an autonomous evaluation criterion for the predefined thing according to the original evaluation information of the reference population for the predefined thing; according to the independent evaluation The standard performs an autonomous evaluation of the target object; and determines an output for the target object based on the self-evaluation.
  • an autonomous evaluation method for an artificial intelligence system comprising: performing an output for a target object; generating an autonomy for the predefined thing according to the original evaluation information of the reference population for the predefined thing An evaluation criterion; autonomous evaluation of the target object according to the self-evaluation criterion; and adjusting the output of the output module for the target object according to the self-evaluation.
  • the system and method provided by the invention construct an autonomous evaluation system of the artificial intelligence system for external things, so that the artificial intelligence system can have the ability to independently evaluate the target object.
  • FIG. 1A is a block diagram of an artificial intelligence system with evaluation capabilities, in accordance with one embodiment of the present invention.
  • FIG. 1B is a block diagram of an artificial intelligence system with evaluation capabilities in accordance with another embodiment of the present invention.
  • FIG. 2 is a block diagram of a storage form of evaluation standard vocabulary information in an evaluation standard vocabulary module according to an embodiment of the present invention
  • FIG. 3 is a flow chart of an artificial intelligence system for determining autonomous evaluation criteria for a thing, in accordance with one embodiment of the present invention
  • FIG. 5 is a flow chart of an artificial intelligence system for determining an autonomous evaluation criteria for a thing, in accordance with another embodiment of the present invention.
  • FIG. 6 is a flow chart showing an update process of an autonomous evaluation criteria for a transaction in an update evaluation library module, in accordance with one embodiment of the present invention
  • FIG. 7 is a flow chart showing an update process of an autonomous evaluation criterion for a thing in an update evaluation library module according to another embodiment of the present invention.
  • FIG. 8 is a flow diagram of another update process for updating an autonomous evaluation criteria for a transaction in an evaluation library module, in accordance with another embodiment of the present invention.
  • FIG. 9 is a flow chart of an artificial intelligence system making an autonomous evaluation of a target object in accordance with an embodiment of the present invention.
  • FIG. 10 is a flow chart of a response output of an artificial intelligence system according to an autonomous evaluation according to an embodiment of the present invention
  • FIG. 11 is a flow chart of a response output of an artificial intelligence system according to an autonomous evaluation according to another embodiment of the present invention.
  • FIG. 12 is a flow chart of a response output of an artificial intelligence system based on an autonomous evaluation in accordance with another embodiment of the present invention.
  • Predefined things refers to things that a predetermined artificial intelligence system needs to determine the criteria for autonomous evaluation.
  • Target thing refers to something that needs artificial intelligence system to evaluate it autonomously. It can be something or a certain piece of music, an expression, etc.
  • Reference Crowd refers to a collection of people referenced in determining an evaluation of a predefined thing.
  • the reference population may be a family member or a collection of people of a certain age group.
  • the original evaluation information refers to information obtained by the perception module from the evaluation subject (the person giving the evaluation) to the evaluation of the predefined thing.
  • Evaluation standard vocabulary information refers to information pre-configured in a library (memory) for evaluation of things summarized according to experience in life, which can be expressed in the form of text, audio, expression images, etc., as a determination of "predefined things" Standard information for independent evaluation.
  • the self-evaluation criterion refers to the self-evaluation of the thing determined by the artificial intelligence system according to the original evaluation information of the reference group for the thing.
  • the expression may be an integer such as a number from 1 to 100, or may be a level, such as a Level, Level 2, Level 3.
  • FIGS 1A and 1B are block diagrams of two evaluation-capable artificial intelligence systems 100 and 100', in accordance with an embodiment of the present invention.
  • the artificial intelligence systems 100 and 100' include an evaluation capability module 101 and a decision module 102, wherein the evaluation capability module 101 is configured to generate an autonomous evaluation criterion for a predefined thing according to the original evaluation information of the reference population for the predefined thing and according to the autonomous evaluation standard. Conduct independent evaluation of the target things.
  • the decision module 102 is configured to determine an output for the target item based on the autonomous evaluation determined by the evaluation capability module 101.
  • the artificial intelligence systems 100 and 100' further include one or more of the following modules, the sensing module 103 configured to obtain original evaluation information of the reference crowd for things, such as the original evaluation information 106 shown in FIG. And identifying the target thing to be self-evaluated; the self-evaluation criterion module 104 is configured to store the self-evaluation criterion made by the evaluation capability module 101; the evaluation standard vocabulary information module 105 is configured to store the experience according to the experience in life. The standard vocabulary information for the evaluation of the thing; and the predefined things module 107 for storing the predetermined things that require the artificial intelligence system to determine the autonomous evaluation criteria.
  • the perception module 103 as shown in FIG.
  • the evaluation capability module 101 is coupled to the autonomous evaluation criteria module 104, and the evaluation criteria vocabulary information module 105 can be coupled to the evaluation capability module 101, and the predefined transaction module 107 may be coupled to the perception module 103 and the evaluation capability module 101, respectively.
  • the self-evaluation criteria module 104 can be coupled to the evaluation capability module 101.
  • the artificial intelligence system 100' differs from the artificial intelligence system 100 of FIG. 1A in that it also includes an output module 113.
  • the output module 113 is for executing an output 114 for the target thing.
  • the output module 113 performs an output 114 for the target object
  • the evaluation capability module 101 is configured to generate an autonomous evaluation criterion for the predefined thing according to the original evaluation information of the reference population for the predefined thing and according to the autonomous evaluation standard. Conduct independent evaluation of the target things.
  • the decision module 102 is configured to adjust the output 114 of the target transaction based on the autonomous evaluation hop determined by the evaluation capability module 101.
  • the energy system 100' adjusts an embodiment of the output 114 of the target object based on the autonomous evaluation determined by the evaluation capability module 101.
  • the other modules of the artificial intelligence system 100' are the same as the artificial intelligence system 100, and can play the same role. For the sake of brevity, no further details are provided herein.
  • the evaluation criteria vocabulary information module 105 can be pre-configured to store evaluation criteria vocabulary information that has become a life experience in social life, and these evaluation criteria vocabulary information can be written , pictures, audio, numbers, etc. exist and are mapped to each other.
  • the evaluation criteria vocabulary information can include a quantitative assessment and map it to corresponding other forms of evaluation criteria vocabulary information.
  • the evaluation may be a number from 0 to 100, with 0 being a very poor evaluation and 100 being an excellent evaluation.
  • the evaluation criteria vocabulary information present in the form of text, pictures, audio, numbers, etc. described above may be mapped to the quantified evaluation.
  • the "good” text 201 above may be mapped to the number 50; the "bad” audio 202 may be mapped to the number 20; the "beautiful" person's portrait photo 203 may be mapped to the number 100.
  • a predefined library of things (not shown) can also be included in the system that can be pre-configured to store things to be evaluated and/or their feature information.
  • different evaluation criteria vocabulary information can be mapped to the same quantitative evaluation, for example, "good” text 201 and "smile” expression 204 can be simultaneously mapped to the quantitative evaluation number "50"
  • the evaluation capability module 101 is configured to set a reference population of autonomous evaluation criteria prior to conducting the evaluation.
  • the reference group can be a family member, a group of people of a certain age, and the like.
  • the smart device 100 will determine an autonomous evaluation criterion for the target thing with reference to the evaluation of the target person by the reference crowd.
  • the awareness module 103 can include an image sensing unit, an audio sensing unit, and other similar physical units that obtain information from the outside world, configured to obtain raw rating information 106 made by a reference population for predefined things in a predefined library of things.
  • the perception module 103 can be configured to obtain an evaluation of the family member's predefined things in the predefined transaction library as the original rating information 106.
  • the evaluation capability module 106 can then match the obtained original rating information 106 with the rating criteria vocabulary information in the rating criteria vocabulary information module 105 and, in turn, determine autonomous evaluation criteria for the predefined thing. For example, referring to FIG. 1 and FIG. 2, the original evaluation information 106 acquired by the sensing module 103 for the transaction A in the predefined transaction library is “bad” audio, and the evaluation capability module 106 compares it with the evaluation standard vocabulary information module 105.
  • the stored evaluation criteria vocabulary information in the match is matched to the "bad" audio 202 in the evaluation criteria vocabulary information module 105, and in turn based on the mapping between the "bad” audio 202 and the number "20"
  • the self-evaluation criterion for the thing A is determined as "20".
  • the price capability module 106 may also determine an autonomous evaluation criteria for predefined things based on the degree of matching between the original rating information 106 and the rating criteria vocabulary information.
  • other parameters or factors can be added to determine the autonomous evaluation criteria based on the degree of matching.
  • FIG. 3 illustrates a flowchart 300 of an artificial intelligence system for determining autonomous evaluation criteria for predefined things, in accordance with one embodiment.
  • the perception module 103 obtains the original rating information 106 of the reference population for the predefined transaction A in the predefined transaction library.
  • the evaluation capability module 101 matches the original evaluation information 106 with the evaluation standard vocabulary information in the evaluation standard vocabulary information module 105.
  • the evaluation capability module 101 determines an autonomous evaluation criterion for the predefined transaction A based on the matching result.
  • FIG. 400 A flowchart 400 of an artificial intelligence system for determining autonomous evaluation criteria for predefined things in accordance with another embodiment is shown in FIG.
  • the evaluation capability module asks a reference population for a predefined transaction A in a predefined transaction library.
  • the question can be presented to the reference population in various forms by an output device external to the smart device, such as a display, speaker, or the like.
  • the question can be an evaluation of the predefined thing A directly to the reference population.
  • the perception module 103 obtains the user's answer to the question.
  • receipt of the answer may be obtained via an image sensing unit or an audio sensing unit or the like of the sensing module 103.
  • the evaluation capability module 101 matches the answer as the original evaluation information 106 with the evaluation criteria vocabulary information in the evaluation criteria vocabulary information module 105.
  • the evaluation capability module 101 determines an autonomous evaluation criterion for the predefined transaction A based on the matching result.
  • FIG. 1 A flow diagram 500 of an artificial intelligence system for determining autonomous evaluation criteria for predefined things in accordance with another embodiment is shown in FIG.
  • the artificial intelligence system 100 further includes a communication module 109 that is communicatively coupled to the network 110 and the sensing module 103 and that communicates information received from the network 110 to Perception module 103.
  • communication module 109 searches via network 110 for network rating information 112 for predefined things A in the predefined transaction library on the network and transmits the network evaluation information 112 to perception module 103.
  • the evaluation capability module 101 compares the network evaluation information 112 with the evaluation standard vocabulary information in the evaluation standard vocabulary information module 105. Make a match.
  • the evaluation capability module 101 determines an autonomous evaluation criterion for the predefined transaction A based on the matching result.
  • the artificial intelligence system further includes an autonomous evaluation criteria module 104.
  • the evaluation capability module 101 determines the autonomous evaluation criteria for the predefined things
  • the autonomous evaluation for the predefined things can be stored to the self-evaluation criteria.
  • Module 104 is used to prepare for subsequent calls.
  • FIG. 6 illustrates an update process 600 for updating an autonomous evaluation criteria for predefined things in the autonomous evaluation criteria module 104, in accordance with one embodiment of the present invention.
  • the perception module 103 obtains re-evaluation information for the predefined transaction A that has been evaluated.
  • the perception module 103 can obtain re-evaluation information for the predefined transaction A according to the process of obtaining the original evaluation information 106 and/or the network evaluation information 112 as explained with reference to FIGS. 3, 5, and 5.
  • the evaluation capability module 101 matches the re-evaluation information with the evaluation criteria vocabulary information in the evaluation criteria vocabulary module 105, and determines a re-autonomous evaluation criterion for the predefined transaction A based on the matching result.
  • the evaluation capability module 101 weights the existing autonomous evaluation criteria for the predefined transaction A in the evaluation library module with the re-autonomous evaluation criteria for the predefined transaction A to obtain a new self-evaluation criterion for the predefined transaction A. To update the existing independent evaluation criteria for the predefined transaction A in the evaluation library module.
  • the autonomous evaluation criteria update process shown in FIG. 6 above may continue to run cyclically, that is, each time an evaluation of a predefined thing is observed, the autonomous evaluation criteria for the predefined thing is updated.
  • the weights in the weighted addition in the above process 600 may be determined according to various factors, for example, may be set such that the existing self-evaluation criteria are equal to the weights of the self-evaluation criteria (for example, both equal to 0.5), or may be based on the individual who made the evaluation.
  • the difference for example, the relationship between family members and strangers
  • the weights of evaluations made by family members may be greater than the weights corresponding to strangers.
  • FIG. 7 An update process 700 for updating the autonomous evaluation criteria for predefined things in the autonomous evaluation criteria criteria module 104 is shown in FIG. 7 in accordance with one embodiment of the present invention.
  • the evaluation capability module divides the reference population into levels while setting the evaluation reference population, for example, into two levels, family members and strangers.
  • the perception module 103 obtains re-evaluation information for the predefined transaction A that has been evaluated.
  • the perception module 103 can be solved according to the figures 3, 4, and 5
  • the process of obtaining the original rating information 106 and/or the network rating information 112 is said to obtain re-evaluation information for the predefined transaction A.
  • the evaluation capability module 101 matches the re-evaluation information with the evaluation criteria vocabulary information in the evaluation criteria vocabulary module 105, and determines a re-self-evaluation criterion for the predefined transaction A based on the matching result.
  • the evaluation capability module 101 assigns the existing self-evaluation criteria for the predefined transaction A in the evaluation library module different from the re-automatic evaluation criteria for the predefined transaction A according to the difference of the reference population level to which the individual making the evaluation belongs. The weight, and then the existing independent evaluation criteria and the re-independent evaluation criteria are weighted according to their respective weights to obtain a new autonomous evaluation criterion for the predefined transaction A to update the existing autonomy for the predefined transaction A in the evaluation library module. evaluation standard.
  • weights may be set by the artificial intelligence system according to other parameters or factors or even randomly within a specific range, so that two The same artificial intelligence system has different evaluations for the same object or thing, so that the artificial intelligence system has considerable independent evaluation standard ability.
  • FIG. 8 illustrates another update process 800 for updating autonomous evaluation criteria for predefined things in the autonomous evaluation criteria module 104 in accordance with another embodiment of the present invention.
  • the evaluation capability module 101 randomly determines autonomous evaluation criteria for the predefined things before determining the autonomous evaluation criteria for the predefined things in the predefined transaction library.
  • the random self-evaluation criterion may be an external representation of the evaluation standard vocabulary information corresponding to the self-evaluation criterion in the evaluation standard vocabulary, for example, displaying the “good” text through an external display, or Play "bad” audio through an external speaker.
  • Such random evaluation can also be expressed directly by means of an independent evaluation standard, for example, an external evaluation standard such as "20" or "40" is displayed through an external display.
  • the perception module 103 receives the response of the reference population to the random assessment in a manner similar to the acquisition of the original assessment information 106 in FIGS.
  • the evaluation capability module modifies the autonomous evaluation criteria for the predefined things based on the response.
  • the self-evaluation criteria for predefined things can be modified by means of updating the self-evaluation criteria similar to those illustrated in FIG. 6 and FIG. 7, that is, the evaluation capability module will be based on the evaluation criteria vocabulary in the vocabulary of the response and evaluation criteria.
  • the matching result of the information determines the re-autonomous evaluation criterion for the random evaluation, and adds the self-evaluation criterion and the randomly-made evaluation weight to modify the self-evaluation criterion for the predefined thing.
  • the evaluation capability module 101 can directly determine the autonomous evaluation criteria corresponding to the response in a manner similar to the manner in which the self-evaluation criteria are determined in FIG. 3 and FIG. 4, and replace the artificial intelligence system with the self-evaluation standard. A random evaluation of the predefined thing.
  • the process of determining and updating autonomous evaluation criteria for predefined things in a predefined transaction library is described above in connection with Figures 2-8. Without departing from the invention, those skilled in the art can think of the evaluation information not in the evaluation standard vocabulary module. If the sensing module often detects the evaluation information in the process of detecting the original evaluation information, the information can be The evaluation standard vocabulary information is inserted into the evaluation standard vocabulary information library module. Further, the corresponding self-evaluation criterion may be inserted into the evaluation standard vocabulary information database, or the user may be required to set an autonomous evaluation criterion corresponding to the evaluation standard vocabulary information.
  • Figure 9 illustrates a process 900 for an artificial intelligence system to make an autonomous evaluation of a target transaction.
  • the artificial intelligence system is triggered to make an autonomous evaluation of the target thing, for example, after receiving an instruction to evaluate the target thing.
  • the perception module 103 identifies the target thing.
  • the evaluation capability module 101 matches the identified target transaction with the thing in the autonomous evaluation criteria module 104 for which the autonomous evaluation criteria have been determined.
  • the evaluation capability module 101 determines an autonomous evaluation of the identified target transaction based on the results of the matching. For example, assume that the autonomous evaluation of the predefined thing A is determined to be "75" in the previous autonomous evaluation determination process as shown in FIGS.
  • the sensing module 103 identifies the target transaction as a predefined transaction A, and the evaluation capability module matches the target transaction with the predefined transaction in the evaluation library module 105, and finds the self-evaluation of the matching result as the predefined transaction A. For "75", the self-evaluation of the target thing is determined as "75".
  • 10 is a flow diagram 1000 of an artificial intelligence system performing a response output based on an autonomous evaluation, in accordance with an embodiment of the present invention.
  • decision module 102 is pre-configured to map each autonomous evaluation to an output corresponding thereto.
  • the evaluation capability module 101 determines an autonomous evaluation of the identified target transaction. For example, the evaluation capability module 101 can determine an autonomous evaluation of the target thing B according to the process of FIG.
  • the decision module determines an output for the identified target object based on the autonomous evaluation of the identified target transaction.
  • the decision module 102 determines, based on the obtained autonomous evaluation and the mapping between the pre-configured autonomous evaluation and the corresponding output, the target object for the identification.
  • the output of B can be performed by the decision module for the target item B, such as the motor-driven artificial intelligence system approaching the target item B.
  • the decision-making module determines the output that the artificial intelligence system should perform, and the self-evaluation of the target object by the artificial intelligence system can be externally expressed.
  • the output here may be a mechanical output close to or away from the target object as described above, or may be a display output of an expression such as "smile” or “tears", and other output expressions may be conceived by those skilled in the art without Deviated from the invention.
  • decision module 102 divides the output into a forward output and a negative output.
  • the evaluation capability module 101 classifies the self-evaluation into a positive parity, a low positive evaluation, a negative evaluation, and a low negative evaluation.
  • the decision module 102 determines the output for the target transaction in accordance with the process provided by FIG. 11, after the output execution unit of the artificial intelligence system is executing the output, the decision module 102 performs the adjustments to the output in accordance with the following steps.
  • step 1102 it is determined whether the output executed at this time is a forward output or a negative output.
  • step 1104 the decision module 102 determines whether the autonomous evaluation of the current target transaction is a positive evaluation, a negative evaluation, or a low negative evaluation, for example, the decision module 102 can communicate with the evaluation capability module 101 to determine the above information. If the autonomous evaluation of the current target transaction is a positive evaluation, then proceeding to step 1108, in step 1108, the decision module 102 determines to enhance the output effect. For example, the enhancement of the output effect may be to drive the artificial intelligence system closer to the target object based on the original motor-driven artificial intelligence system approaching the target object.
  • the decision module 102 determines to take and present Output the opposite reverse output. For example, the current output is that the motor-driven artificial intelligence system is close to the target, and after the decision module 102 determines that the reverse output is taken, the current output can be changed to the motor-driven artificial intelligence system away from the target. If the autonomous evaluation of the current target transaction is a low negative evaluation, then proceeding to step 1110, in step 1110, the decision module 102 determines to attenuate the output effect of the current output. For example, when the current output is a motor-driven artificial intelligence system approaching a target transaction, decision module 102 may decide to approach the target thing more slowly than the current speed.
  • step 1106 the decision module 102 determines whether the autonomous evaluation of the current target object is a positive evaluation, a negative evaluation, or a low positive evaluation, and if it is a negative evaluation, proceeds to step 1108 if it is a low positive direction. Evaluation, proceed to step 1110, if it is positive evaluation, proceed to 1112
  • the step of determining whether the output action is a positive output or a negative output and the step of determining whether the autonomous evaluation of the current target object is a positive evaluation or a negative evaluation may change the order while maintaining the current same technical effect.
  • the output action is a positive output
  • the current self-evaluation is a forward output, and those skilled in the art can think of maintaining the current output effect or enhancing the current output effect without departing from the technology of the present invention. Also.
  • FIG. 12 is a flow diagram 1200 of an artificial intelligence system performing a response output based on an autonomous evaluation, in accordance with another embodiment of the present invention. 1B and 10, the process of FIG. 12 begins at 1201, the artificial intelligence system has executed an output for the target transaction, for example, the output module 113 has performed an output 114 for the target transaction, for example, the output 114 may be Close to the target thing.
  • the output of the decision may be an autonomous evaluation not made based on the artificial intelligence system, such as may be a decision made by the user to control the artificial intelligence system or by the user-specified decision module 102.
  • decision module 102 divides the output into a forward output and a negative output.
  • the evaluation capability module 101 classifies the evaluation into a positive parity, a low positive evaluation, a negative evaluation, and a low negative evaluation.
  • decision module 102 performs the adjustments to the output in accordance with the following steps.
  • step 1202 it is determined whether the output 114 executed at this time is a forward output or a negative output. If it is a forward output, the process proceeds to step 1204, and if it is a negative output, the process proceeds to step 1206.
  • the decision module 102 determines whether the autonomous evaluation of the current target transaction is a positive evaluation, a negative evaluation, or a low negative evaluation, for example, the decision module 102 can communicate with the evaluation capability module 101 to determine the above information.
  • the decision module 102 determines the effect of the enhanced output 114.
  • the enhancement of the output effect may be to drive the artificial intelligence system closer to the target object based on the original motor-driven artificial intelligence system approaching the target object.
  • the decision module 102 determines to take the reverse output opposite the current output 114.
  • the current output 114 is a motor-driven artificial intelligence system that is close to the target, and after the decision module 102 determines that the reverse output is taken, the current output 114 can be changed to a motor-driven artificial intelligence system away from the target.
  • step 1210 the decision module 102 determines to attenuate the output effect of the current output 114. For example, when the current output 114 is a motor-driven artificial intelligence system approaching a target transaction, the decision module 102 can decide to approach the target thing more slowly than the current speed. Similarly, in step 1206, the decision module 102 determines whether the autonomous evaluation of the current target transaction is a positive evaluation, a negative evaluation, or a low positive evaluation, and if it is a negative evaluation, proceeds to step 1208, if it is a low positive The evaluation proceeds to step 1210, and if it is a positive evaluation, proceeds to 1214. I will not repeat them here.
  • the step of determining whether the output action is a positive output or a negative output and the step of determining whether the autonomous evaluation of the current target object is a positive evaluation or a negative evaluation may change the order while maintaining the current same technical effect.
  • the output action is a positive output
  • the current self-evaluation is a forward output, and those skilled in the art can think of maintaining the current output effect or enhancing the current output effect without departing from the technology of the present invention. Also.
  • information, signals, and data may be represented using any of a variety of different technologies and techniques.
  • the data, instructions, commands, information, signals, bits (bits), symbols, and chips referenced throughout the above description may be by voltage, current, electromagnetic waves, magnetic fields or magnetic particles, light fields or optical particles, or any thereof. Combined to represent.
  • DSPs digital signal processors
  • ASICs application specific integrated circuits
  • FPGAs field programmable gate arrays
  • a general purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine.
  • the processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration.
  • a software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
  • An exemplary storage medium is coupled to the processor to enable the processor to read and write information to/from the storage medium.
  • the storage medium can be integrated into the processor.
  • the processor and the storage medium can reside in an ASIC.
  • the ASIC can reside in the user terminal.
  • the processor and the storage medium may reside as a discrete component in the user terminal.
  • the functions described may be implemented in hardware, software, firmware, or any combination thereof. If implemented as a computer program product in software, the functions may be stored on or transmitted as one or more instructions or code on a computer readable medium.
  • Computer readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another.
  • a storage medium may be any available media that can be accessed by a computer.
  • such computer readable media may comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage device, or can be used to carry or store instructions or data structures. Any other medium that is desirable for program code and that can be accessed by a computer.
  • any connection is also properly referred to as a computer readable medium.
  • the software is transmitted from a web site, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave.
  • the coaxial cable, fiber optic cable, twisted pair cable, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of the medium.
  • Disks and discs as used herein include compact discs (CDs), laser discs, optical discs, digital versatile discs (DVDs), floppy discs, and Blu-ray discs, in which disks are often reproduced magnetically. Data, and discs optically reproduce data with a laser. Combinations of the above should also be included within the scope of computer readable media.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Theoretical Computer Science (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Quality & Reliability (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • Marketing (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种具备评价能力的人工智能系统及其评价方法，该系统包括评价能力模块(101)，其配置成根据参考人群对于预定义事物的原始评价信息生成对于该预定义事物的自主评价标准；其还配置成根据该自主评价标准对目标事物进行自主评价；以及决策模块(102)，其配置成根据该自主评价确定针对该目标事物的输出。该方法包括据参考人群对于预定义事物的原始评价信息生成对于该预定义事物的自主评价标准；根据该自主评价标准对目标事物进行自主评价；以及根据该自主评价确定针对该目标事物的输出。

Description

具备评价能力的人工智能系统及其评价方法 技术领域
本发明涉及人工智能领域，尤其涉及具备评价能力的人工智能系统，以及该人工智能系统对于目标事物进行评价的方法。
背景技术
人工智能是研究用于模拟、延伸和扩展人的智能的理论、方法、技术以及应用系统的一门新的技术科学。自从诞生以来，人工智能的理论和技术日益成熟，应用领域不断扩大，作为计算机科学的一个分支，其企图了解智能的实质，并产生出一种新的能以人类智能相似的方式作出反应的智能机器。
现有的人工智能技术的实现，如何使得人工智能机器获得“自我”意识而非按照人类所设定的工序或步骤进行动作在技术思想方面一直面临的巨大挑战。现有的人工智能技术中，对于某个事物的评价是由用户或者生产厂商在预先设置完成，并不能够随着生活经验的不断增加得到不断的更新与补充。也就是说，现有的技术中，人工智能机器所具有的是“被赋予”的意识而非“自我”意识。而具备“自我”意识的前提是能够遵循由经验中形成的评判标准对外界事物做出本身的主观的评价。
本发明提供了一种具备评价能力的人工智能系统及其相关的评价方法，构建了人工智能系统对于外界事物的评价体系，使得人工智能系统能够随着生活经验的增加而具备自主评价目标事物的能力。
发明内容
以下给出一个或多个方面的简要概述以提供对这些方面的基本理解。此概述不是所有构想到的方面的详尽综览，并且既非旨在指认出所有方面的关键性或决定性要素亦非试图界定任何或所有方面的范围。其唯一的目的是要以简化形式给出一个或多个方面的一些概念以为稍后给出的更加详细的描述之序。
在本发明的一个方面，提供了一种具备自主评价能力的人工智能系统，包括：评价能力模块，其配置成根据参考人群对于预定义事物的原始评价信息生成对于该 预定义事物的自主评价标准，该评价能力模块还配置成根据该自主评价标准对目标事物进行自主评价；以及决策模块，其配置成根据该自主评价确定针对该目标事物的输出。
在本发明的另一个方面，提供了另一种具备自主评价能力的人工智能系统，包括：输出模块，其配置成执行对于目标事物的输出；评价能力模块，其配置成根据参考人群对于预定义事物的原始评价信息生成对于该预定义事物的自主评价标准，该评价能力模块还配置成根据该自主评价标准对目标事物进行自主评价；以及决策模块，其配置成根据该自主评价调整该输出模块对于该目标事物的输出。
在本发明的另一个方面，提供了一种用于人工智能系统的自主评价方法，包括：根据参考人群对于预定义事物的原始评价信息生成对于该预定义事物的自主评价标准；根据该自主评价标准对目标事物进行自主评价；以及根据该自主评价确定针对该目标事物的输出。
在本发明的另一个方面，提供了一种用于人工智能系统的自主评价方法，包括：执行对于目标事物的输出；根据参考人群对于预定义事物的原始评价信息生成对于该预定义事物的自主评价标准；根据该自主评价标准对目标事物进行自主评价；以及根据该自主评价调整该输出模块对于该目标事物的该输出。
本发明所提供的系统和方法构建了人工智能系统对于外界事物的自主评价体系，使得人工智能系统能够具备自主评价目标事物的能力。
附图说明
在结合以下附图阅读本公开的实施例的详细描述之后，能够更好地理解本发明的上述特征和优点。在附图中，各组件不一定是按比例绘制，并且具有类似的相关特性或特征的组件可能具有相同或相近的附图标记。
图1A是根据本发明一个实施例的具备评价能力的人工智能系统的框图；
图1B是根据本发明另一个实施例的具备评价能力的人工智能系统的框图；
图2是根据本发明的一个实施例的评价标准词汇库模块中评价标准词汇信息的储存形式的框图；
图3是根据本发明一个实施例的人工智能系统用于确定对于事物的自主评价标准的流程图；
图4是根据本发明另一个实施例的人工智能系统用于确定对于事物的自主评价 标准的流程图；
图5是根据本发明另一个实施例的人工智能系统用于确定对于事物的自主评价标准的流程图；
图6是根据本发明的一个实施例的更新评价库模块中的对于事物的自主评价标准的更新过程流程图；
图7是根据本发明的另一个实施例的更新评价库模块中的对于事物的自主评价标准的更新过程流程图；
图8是根据本发明的另一个实施例的更新评价库模块中的对于事物的自主评价标准的另一更新过程的流程图；
图9是根据本发明的一个实施例的人工智能系统对于目标事物做出自主评价的流程图；
图10是根据本发明的一个实施例的人工智能系统根据自主评价进行响应输出的流程图；
图11是根据本发明的另一个实施例的人工智能系统根据自主评价进行响应输出的流程图；
图12是根据本发明的另一个实施例的人工智能系统根据自主评价进行响应输出的流程图。
具体实施方式
术语说明
首先，出于使得读者更清楚明白地理解本发明的具体实施例的目的，以下对于本文中所采用的数据进行说明：
预定义事物：是指预先确定的人工智能系统要为其确定自主评价标准的事物。
目标事物：是指需要人工智能系统对其进行自主评价的事物，其可以是某个事物也可以是某段音乐、某个表情等。
参考人群：是指在确定对于预定义事物的评价时所参考的人的集合，例如，参考人群可以是家庭成员，也可以是某一段年龄段的人的集合。
原始评价信息：是指感知模块所获取的由评价主体(给予评价的人)对预定义事物的评价的信息。
评价标准词汇信息：是指预先配置在库(存储器)中的根据生活中的经验所总 结的对于事物的评价的信息，其可以文本、音频、表情图像等形式表现，作为确定“预定义事物”的自主评价的标准信息。
自主评价标准：是指人工智能系统根据参考人群对于事物的原始评价信息而确定的对于该事物的自主评价，具体地，其表现形式可以是数字如1-100的整数，也可以是等级，如一级、二级、三级。
本发明的具体实施方式
以下结合附图和具体实施例对本发明作详细描述。注意，以下结合附图和具体实施例描述的诸方面仅是示例性的，而不应被理解为对本发明的保护范围进行任何限制。
参见图1A和1B，图1A和1B是根据本发明的实施例的两个具备评价能力的人工智能系统100和100’的框图。人工智能系统100和100’包括评价能力模块101与决策模块102，其中评价能力模块101用于根据参考人群对于预定义事物的原始评价信息生成对于预定义事物的自主评价标准并且根据该自主评价标准对目标事物进行自主评价。决策模块102用于根据评价能力模块101确定的自主评价确定针对该目标事物的输出。此外，人工智能系统100和100’还包括以下模块中的一者或多者，感知模块103，其配置成获取参考人群对于事物的原始评价信息，例如图1中所示的原始评价信息106，以及识别要对其进行自主评价的目标事物；自主评价标准模块104，用于存储评价能力模块101所做出的自主评价标准；评价标准词汇信息模块105，用于存储根据生活中的经验所总结的对于事物的评价标准词汇信息；以及预定义事物模块107，用于存储预先设定的要求人工智能系统确定自主评价标准的事物。其中，如图1中所示的感知模块103耦合到评价能力模块101，评价能力模块101耦合到自主评价标准模块104，评价标准词汇信息模块105可以耦合到评价能力模块101，而预定义事物模块107可以分别耦合到感知模块103以及评价能力模块101。自主评价标准模块104可以与评价能力模块101耦合。
人工智能系统100’与图1A中的人工智能系统100的差别在于，还包括输出模块113。输出模块113用于执行针对目标事物的输出114。在人工智能系统100’中输出模块113执行针对目标事物的输出114，评价能力模块101用于根据参考人群对于预定义事物的原始评价信息生成对于预定义事物的自主评价标准并且根据该自主评价标准对目标事物进行自主评价。决策模块102用于根据评价能力模块101确定的自主评价跳调整对该目标事物的输出114。在下文中将结合图12解说人工智 能系统100’根据评价能力模块101确定的自主评价调整对该目标事物的输出114的实施例。人工智能系统100’的其他模块与人工智能系统100相同，并可以起到相同的作用，为了简要起见，此处不再赘述。
在一个方面，如图2中所示，在进行自主评价之前，评价标准词汇信息模块105可以被预先配置成存储在社会生活中已经成为生活经验的评价标准词汇信息，这些评价标准词汇信息可以文字、图片、音频、数字等形式存在，并且相互映射。例如“好”的文字201、“不好”音频202、被认为是“美”的人的肖像照203、微笑的表情图片204等。在进一步的方面，评价标准词汇信息可以包括量化的评价并将其与相应的其他形式的评价标准词汇信息映射。例如，评价可以是0-100的数字，0为极差的评价而100为极好的评价。在一个方面，上文中所述的文字、图片、音频、数字等形式存在的评价标准词汇信息可以与量化的评价进行映射。例如，上文中“好”的文本201可以被映射到数字50；“不好”的音频202可以被映射到数字20,；“美”的人的肖像照203可以被映射到数字100。在另一个方面，系统中还可包括预定义事物库(未示出)，其可以被预先配置成存储所要评价的事物和/或其特征信息。此外，不同的评价标准词汇信息可以映射到相同的量化的评价，例如“好”的文本201和“微笑”的表情204可以同时映射到量化评价数字“50”
继续参见图1，评价能力模块101被配置成在进行评价之前设置自主评价标准的参考人群。该参考人群可以是家庭成员，也可以是某一年龄段的人群等等。智能设备100将参照参考人群对于目标事物的评价而确定对目标事物的自主评价标准。感知模块103可以包括图像感知单元、音频感知单元以及其他类似的从外界获取信息的物理单元，其被配置成获取参考人群对于预定义事物库中的预定义事物所做出的原始评价信息106。例如，当参考人群被配置成家庭成员时，感知模块103可以被配置成获取家庭成员对于预定义事物库中的预定义事物所做出的评价，将其作为原始评价信息106。评价能力模块106可以随后将所获得的原始评价信息106与评价标准词汇信息模块105中的评价标准词汇信息匹配，并进而确定对于该预定义事物的自主评价标准。例如，参照图1与图2，感知模块103所获取的对于预定义事物库中的事物A的原始评价信息106为“不好”的音频，评价能力模块106将其与评价标准词汇信息模块105中的所存储的评价标准词汇信息进行匹配，其与评价标准词汇信息模块105中的“不好”的音频202匹配，并进而根据“不好”的音频202与数字“20”之间的映射将对于该事物A的自主评价标准确定为“20”。当然，评 价能力模块106也可以根据原始评价信息106与评价标准词汇信息之间的匹配度来确定对于预定义事物的自主评价标准。例如，当感知模块103获取到某人物B的照片时，评价能力模块106将其与评价标准词汇信息中的肖像照203进行匹配，确定该两者之间的匹配度为85％，则评价能力模块106可以根据该匹配度85％以及肖像照203与数字“100”之间的映射关系来确定人物B的自主评价标准，例如但不限于人物B的自主评价标准为100*85％＝85。当然其他的参数或因数可以被增加以根据匹配度来确定自主评价标准。
例如，图3示出了根据一个实施例的人工智能系统用于确定预定义事物的自主评价标准的流程图300。结合图1以及图2，在步骤302，感知模块103获取参考人群对于预定义事物库中的预定义事物A的原始评价信息106。在步骤304，评价能力模块101将原始评价信息106与评价标准词汇信息模块105中的评价标准词汇信息进行匹配。在步骤306，评价能力模块101根据匹配结果确定预定义事物A的自主评价标准。
图4中示出了根据另一个实施例的人工智能系统用于确定预定义事物的自主评价标准的流程图400。结合图1以及图2，在步骤402，评价能力模块针对预定义事物库中的预定义事物A向参考人群提问。在一个方面该提问可以通过智能设备外接的输出设备，诸如显示器、扬声器等以各种形式向参考人群提出。在另一个方面，该提问可以是直接针对参考人群对于预定义事物A的评价的。在步骤404，感知模块103获取用户对于该提问的回答。作为示例而非限定，该回答的接收可以经由感知模块103的图像感知单元或音频感知单元等获得。在步骤406，类似于参考图3所描述的步骤304，评价能力模块101将该回答作为原始评价信息106与评价标准词汇信息模块105中的评价标准词汇信息进行匹配。在步骤408，类似于步骤306，评价能力模块101根据匹配结果确定预定义事物A的自主评价标准。
图5中示出了根据另一个实施例的人工智能系统用于确定预定义事物的自主评价标准的流程图500。如图1中所示的，在本发明的一个实施例中，人工智能系统100还包括通信模块109，其能与网络110以及感知模块103通信耦合，并且将从网络110接收到的信息传送给感知模块103。在流程图500中，在步骤502，通信模块109经由网络110搜索网络上对于预定义事物库中的预定义事物A的网络评价信息112，并将该网络评价信息112传送给感知模块103。在步骤504，评价能力模块101将该网络评价信息112与评价标准词汇信息模块105中的评价标准词汇信息 进行匹配。在步骤506，类似于步骤306，评价能力模块101根据匹配结果确定预定义事物A的自主评价标准。
以上三种确定对于预定义事物的自主评价标准的流程仅作为示例而非限定，本领域普通技术人员可以当然地根据本发明的教导而想到人工智能系统可以之对于预定义事物做出自主评价标准的流程。如图1中所示的，人工智能系统还包括自主评价标准模块104，在评价能力模块101确定了对于预定义事物的自主评价标准之后，可以将对于预定义事物的自主评价存储到自主评价标准模块104之中以备之后的调用。
图6示出了根据本发明的一个实施例的更新自主评价标准模块104中的对于预定义事物的自主评价标准的更新过程600。在步骤602中，感知模块103获取对于已经评价过的预定义事物A的重新评价信息。例如，感知模块103可以根据参照图3、5、5所解说的获取原始评价信息106和/或网络评价信息112的流程来获取对于预定义事物A的重新评价信息。在步骤604，评价能力模块101将该重新评价信息与评价标准词汇库模块105中的评价标准词汇信息匹配，并且根据匹配结果确定对于预定义事物A的再次自主评价标准。在步骤606，评价能力模块101将评价库模块中对于预定义事物A的现有自主评价标准与对于预定义事物A的再次自主评价标准加权相加以获得对于预定义事物A的新的自主评价标准以更新评价库模块中对于预定义事物A的现有自主评价标准。
对于每一个预定义事物，上述图6中所示的自主评价标准更新过程可以持续循环运行，即每观察到一次对于预定义事物的评价就更新对于该预定义事物的自主评价标准。上述过程600中的加权相加中的权重可以根据多种因素确定，例如可以设置成现有自主评价标准与再次自主评价标准的权重相等(例如都等于0.5)，也可以根据做出评价的个体的不同(例如，家庭成员与陌生人的关系)而对不同的自主评价标准施加不同的权重，例如家庭成员所做出的评价所对应的权重可以大于陌生人所对应的权重。
图7中根据本发明的一个实施例示出了更新自主评价标准标准模块104中的对于预定义事物的自主评价标准的更新过程700。在步骤702，评价能力模块在设置评价参考人群的同时将参考人群划分成等级，例如分为两个等级，家庭成员与陌生人。在步骤704，类似于图6中的步骤602，感知模块103获取对于已经评价过的预定义事物A的重新评价信息。例如，感知模块103可以根据参照图3、4、5所解 说的获取原始评价信息106和/或网络评价信息112的流程来获取对于预定义事物A的重新评价信息。在步骤706，评价能力模块101将该重新评价信息与评价标准词汇库模块105中的评价标准词汇信息匹配，并且根据匹配结果确定对于预定义事物A的再次自主评价标准。在步骤708，评价能力模块101根据做出评价的个体所属的参考人群等级的不同赋予评价库模块中对于预定义事物A的现有自主评价标准与对于预定义事物A的再次自主评价标准不同的权重，进而将现有自主评价标准与再次自主评价标准根据其各自不同的权重加权相加以获得对于预定义事物A的新的自主评价标准以更新评价库模块中对于预定义事物A的现有自主评价标准。
在步骤708中，由于做出原始评价信息的主体所述的参考人群的级别不同赋予现有自主评价标准以及再次自主评价标准以不同的权重，例如，现有自主评价标准所基于的原始评价信息是由陌生人所做出的，赋予权重0.3，而再次自主评价标准所给予的原始评价信息是由家庭成员所做出的，则赋予权重0.7。那么新的自主评价标准＝0.3*现有自主评价标准+0.7*再次自主评价标准。由此可以看出，通过使用图7中所用的更新过程700，可以自由设定不同参考人群对于人工智能系统的自主评价标准的影响力，在上述示例中，家庭成员对于最终的自主评价标准的影响力要高于陌生人。
上述示例仅作为示例而非限定，其他参数或者因素可以被加入加权相加的过程中，甚至与权重可以由人工智能系统根据其他参数或因素甚至随机在一个具体范围内设定，这样可以使得两个相同的人工智能系统对于同一对象或事物由不同的评价，使得人工智能系统具有相当的自主评价标准能力。
图8示出了根据本发明另一个实施例更新自主评价标准模块104中的对于预定义事物的自主评价标准的另一更新过程800。
在步骤802中，评价能力模块101在确定对于预定义事物库中的预定义事物的自主评价标准之前，随机确定对于预定义事物的自主评价标准。例如这种随机自主评价标准可以是以评价标准词汇库中的与该自主评价标准相对应的评价标准词汇信息为外在表现形式的，例如将“好”的文字通过外接的显示器显示出来，或者将“不好”的音频通过外接的扬声器播放出来。这种随机评价也可以是直接通过自主评价标准的方式表现出来，例如通过外接的显示器显示“20”、“40”等自主评价标准。在步骤804，感知模块103以类似于图3、图4中获取原始评价信息106的方式接收参考人群对于该随机评价的反应，例如，通过图像感知单元或者音频感知 单元获取参考人群的对于该随机评价的反应。该反应的形式可以类似于图3、图4中的原始评价信息106，例如“好”的文字输入、“不好”的音频等。在步骤806，评价能力模块根据该反应来修改对于预定义事物的自主评价标准。例如，可以通过类似于图6、图7中所解说的更新自主评价标准的方式来修改对于预定义事物的自主评价标准，即评价能力模块将根据该反应与评价标准词汇库中的评价标准词汇信息的匹配结果确定对于该随机评价的再次自主评价标准，将再次自主评价标准与随机做出的评价加权相加来修改对于预定义事物的自主评价标准。在例如，评价能力模块101可以直接通过与图3、图4中确定自主评价标准的方式类似的方式来确定该反应所对应的自主评价标准，用该自主评价标准替代人工智能系统所做出的对于该预定义事物的随机评价。
以上结合图2-图8描述了确定以及更新对于预定义事物库中预定义事物的自主评价标准的过程。在不背离本发明的前提下，本领域技术人员可以想到对于不在评价标准词汇库模块中的评价信息，如果感知模块经常在检测原始评价信息的过程中检测到该评价信息，那么可以将该信息作为评价标准词汇信息插入评价标准词汇信息库模块中。进一步地，可以将其对应的自主评价标准插入评价标准词汇信息库中，或者也可以要求用户设置对应于该评价标准词汇信息的自主评价标准。
如以上所述的，对于预定义事物的自主评价被存储在评价库中以备将来调用。图9示出了人工智能系统对于目标事物做出自主评价的流程900。当人工智能系统被触发对目标事物做出自主评价后，例如接收到对于目标事物做出评价的指令后。在步骤902，感知模块103识别目标事物。在步骤904，评价能力模块101将所识别到的目标事物与自主评价标准模块104中的已经为其确定自主评价标准的事物进行匹配。在步骤906，评价能力模块101根据匹配的结果确定对于所识别的目标事物的自主评价。例如，假设在之前的如图3、4、5中的自主评价确定过程中将预定义事物A的自主评价确定为“75”，当图1中的人工智能系统100接收到对目标事物做出评价的指令后其感知模块103识别该目标事物为预定义事物A，评价能力模块将该目标事物与评价库模块105中的预定义事物进行匹配，查找到匹配结果为预定义事物A的自主评价为“75”，则该目标事物的自主评价则被确定为“75”。当然，其他的匹配方式也可以被应用。例如评价能力模块经过匹配认为感知模块103所识别到目标事物与预定义事物A的匹配度为80％，而预定义事物A的自主评价为“75”，则目标事物的自主评价可以被确定为75*0.8＝60。以上所述的实施例仅 作为示例而非限定，其他的匹配方式以及因素也可以被本领域技术人员采用或应用而不背离本发明。
图10是根据本发明的一个实施例的人工智能系统根据自主评价进行响应输出的流程图1000。结合图10，决策模块102在预先被配置成将每个自主评价映射到与之相应的输出。在步骤1002，评价能力模块101确定了识别的目标事物的自主评价。例如，评价能力模块101可以根据图9的过程确定对于目标事物B的自主评价。在步骤1004，决策模块根据所识别的目标事物的自主评价确定针对所识别的目标事物的输出。例如，决策模块102在从评价能力模块101获取到目标事物B的自主评价之后，决策模块根据所获得的自主评价以及预先配置的自主评价与相应输出之间的映射确定针对该所识别的目标事物B的输出。再进一步的方面，人工智能系统的输出执行机构(未示出)可以根据决策模块所确定的针对目标事物B的输出来执行该输出，例如电动机驱动人工智能系统靠近目标事物B。
按照图10所提供的过程，通过决策模块决策人工智能系统所应该执行的输出，可以将人工智能系统对于目标事物所进行的自主评价外在地表现出来。这里的输出可以是如上所述的靠近或远离目标物体之类的机械输出，也可以是“微笑”“流泪”等表情的显示输出，本领域普通技术人员还可以想到其他的输出表现形式而不背离本发明。
图11是根据本发明的另一个实施例的人工智能系统根据自主评价进行响应输出的流程图1100。结合图1，决策模块102将输出分为正向输出以及负向输出。评价能力模块101将自主评价分为正向平价、低正向评价、负向评价以及低负向评价。在决策模块102按照图11所提供的过程决定针对目标事物的输出，人工智能系统的输出执行单元正在执行该输出后，决策模块102按照以下步骤执行对于输出的调整。在步骤1102，判断此时所执行的输出是正向输出还是负向输出，若为正向输出则前进到步骤1104，若为负向输出则前进到步骤1106。在步骤1104，决策模块102确定当前目标事物的自主评价是为正向评价、负向评价、还是低负向评价，例如决策模块102可以与评价能力模块101通信以确定上述信息。若当前目标事物的自主评价为正向评价则前进到步骤1108，在步骤1108中，决策模块102确定加强输出效果。例如，该输出效果的加强可以是在原先电动机驱动人工智能系统靠近目标事物的基础上驱动人工智能系统更快速地靠近目标事物。若当前目标事物的自主评价为负向评价，则前进到步骤1112，在步骤1112中，决策模块102确定采取与当前 输出相反的反向输出。例如，当前的输出为电动机驱动人工智能系统靠近目标事物，而决策模块102确定采取反向输出之后，可以将当前输出更改为电动机驱动人工智能系统远离目标事物。若当前目标事物的自主评价为低负向评价，则前进到步骤1110，在步骤1110中，决策模块102确定减弱当前输出的输出效果。例如，当前的输出为电动机驱动人工智能系统靠近目标事物时，决策模块102可以决定比当前的速度更慢速地靠近目标事物。类似地，在步骤1106中，决策模块102确定当前目标事物的自主评价是为正向评价、负向评价还是低正向评价，若为负向评价，则前进到步骤1108，若为低正向评价，则前进到步骤1110，若为正向评价则前进到1112
以上的实施例仅作为示例，本领域技术人员在不背离本发明的基础上可以对上述实施例中的步骤进行步骤的调换以及补充。例如，确定输出动作为正向输出还是负向输出的步骤以及确定当前目标事物的自主评价是正向评价还是负向评价的步骤可以调换顺序，而保持当前相同的技术效果。再例如，若输出动作为正向输出，当前的自主评价是若正向输出，本领域普通技术人员在不背离本发明的技术上可以想到保持当前的输出效果，或者加强当前的输出效果，反之亦然。
图12是根据本发明的另一个实施例的人工智能系统根据自主评价进行响应输出的流程图1200。结合图1B和图10，图12的过程开始于1201，人工智能系统已经执行了针对于目标事物的某项输出，例如输出模块113已经执行了针对目标事物的输出114，例如该输出114可以是靠近目标事物。注意，该决定的输出可以是不基于人工智能系统所做出的自主评价，例如可以是由用户控制人工智能系统的或者由用户指定决策模块102所做出的决策。同样结合图10，决策模块102将输出分为正向输出以及负向输出。评价能力模块101将评价分为正向平价、低正向评价、负向评价以及低负向评价。在1201之后，决策模块102按照以下步骤执行对于输出的调整。在步骤1202，判断此时所执行的输出114是正向输出还是负向输出，若为正向输出则前进到步骤1204，若为负向输出则前进到步骤1206。在步骤1204，决策模块102确定当前目标事物的自主评价是为正向评价、负向评价、还是低负向评价，例如决策模块102可以与评价能力模块101通信以确定上述信息。若当前目标事物的自主评价为正向评价则前进到步骤1208，在步骤110中，决策模块102确定加强输出114的效果。例如，该输出效果的加强可以是在原先电动机驱动人工智能系统靠近目标事物的基础上驱动人工智能系统更快速地靠近目标事物。若当前目标 事物的自主评价为负向评价，则前进到步骤1212，在步骤1212中，决策模块102确定采取与当前输出114相反的反向输出。例如，当前的输出114为电动机驱动人工智能系统靠近目标事物，而决策模块102确定采取反向输出之后，可以将当前输出114更改为电动机驱动人工智能系统远离目标事物。若当前目标事物的自主评价为低负向评价，则前进到步骤1210，在步骤1210中，决策模块102确定减弱当前输出114的输出效果。例如，当前的输出114为电动机驱动人工智能系统靠近目标事物时，决策模块102可以决定比当前的速度更慢速地靠近目标事物。类似地，在步骤1206中，决策模块102确定当前目标事物的自主评价是为正向评价、负向评价还是低正向评价，若为负向评价，则前进到步骤1208，若为低正向评价，则前进到步骤1210，若为正向评价则前进到1214。此处不再赘述。
以上的实施例仅作为示例，本领域技术人员在不背离本发明的基础上可以对上述实施例中的步骤进行步骤的调换以及补充。例如，确定输出动作为正向输出还是负向输出的步骤以及确定当前目标事物的自主评价是正向评价还是负向评价的步骤可以调换顺序，而保持当前相同的技术效果。再例如，若输出动作为正向输出，当前的自主评价是若正向输出，本领域普通技术人员在不背离本发明的技术上可以想到保持当前的输出效果，或者加强当前的输出效果，反之亦然。
尽管为使解释简单化将上述方法图示并描述为一系列动作，但是应理解并领会，这些方法不受动作的次序所限，因为根据一个或多个实施例，一些动作可按不同次序发生和/或与来自本文中图示和描述或本文中未图示和描述但本领域技术人员可以理解的其他动作并发地发生。
本领域技术人员将可理解，信息、信号和数据可使用各种不同技术和技艺中的任何技术和技艺来表示。例如，以上描述通篇引述的数据、指令、命令、信息、信号、位(比特)、码元、和码片可由电压、电流、电磁波、磁场或磁粒子、光场或光学粒子、或其任何组合来表示。
本领域技术人员将进一步领会，结合本文中所公开的实施例来描述的各种解说性逻辑板块、模块、电路、和算法步骤可实现为电子硬件、计算机软件、或这两者的组合。为清楚地解说硬件与软件的这一可互换性，各种解说性组件、框、模块、电路、和步骤在上面是以其功能性的形式作一般化描述的。此类功能性是被实现为硬件还是软件取决于具体应用和施加于整体设备的设计约束。技术人员对于每种特定应用可用不同的方式来实现所描述的功能性，但这样的实现决策不应被解读成导 致脱离了本发明的范围。
结合本文所公开的实施例描述的各种解说性逻辑模块、和电路可用通用处理器、数字信号处理器(DSP)、专用集成电路(ASIC)、现场可编程门阵列(FPGA)或其它可编程逻辑器件、分立的门或晶体管逻辑、分立的硬件组件、或其设计成执行本文所描述功能的任何组合来实现或执行。通用处理器可以是微处理器，但在替换方案中，该处理器可以是任何常规的处理器、控制器、微控制器、或状态机。处理器还可以被实现为计算设备的组合，例如DSP与微处理器的组合、多个微处理器、与DSP核心协作的一个或多个微处理器、或任何其他此类配置。
结合本文中公开的实施例描述的方法或算法的步骤可直接在硬件中、在由处理器执行的软件模块中、或在这两者的组合中体现。软件模块可驻留在RAM存储器、闪存、ROM存储器、EPROM存储器、EEPROM存储器、寄存器、硬盘、可移动盘、CD-ROM、或本领域中所知的任何其他形式的存储介质中。示例性存储介质耦合到处理器以使得该处理器能从/向该存储介质读取和写入信息。在替换方案中，存储介质可以被整合到处理器。处理器和存储介质可驻留在ASIC中。ASIC可驻留在用户终端中。在替换方案中，处理器和存储介质可作为分立组件驻留在用户终端中。
在一个或多个示例性实施例中，所描述的功能可在硬件、软件、固件或其任何组合中实现。如果在软件中实现为计算机程序产品，则各功能可以作为一条或更多条指令或代码存储在计算机可读介质上或藉其进行传送。计算机可读介质包括计算机存储介质和通信介质两者，其包括促成计算机程序从一地向另一地转移的任何介质。存储介质可以是能被计算机访问的任何可用介质。作为示例而非限定，这样的计算机可读介质可包括RAM、ROM、EEPROM、CD-ROM或其它光盘存储、磁盘存储或其它磁存储设备、或能被用来携带或存储指令或数据结构形式的合意程序代码且能被计算机访问的任何其它介质。任何连接也被正当地称为计算机可读介质。例如，如果软件是使用同轴电缆、光纤电缆、双绞线、数字订户线(DSL)、或诸如红外、无线电、以及微波之类的无线技术从web网站、服务器、或其它远程源传送而来，则该同轴电缆、光纤电缆、双绞线、DSL、或诸如红外、无线电、以及微波之类的无线技术就被包括在介质的定义之中。如本文中所使用的盘(disk)和碟(disc)包括压缩碟(CD)、激光碟、光碟、数字多用碟(DVD)、软盘和蓝光碟，其中盘(disk)往往以磁的方式再现数据，而碟(disc)用激光以光学方式再现数据。上述的组合也应被包括在计算机可读介质的范围内。
提供对本公开的先前描述是为使得本领域任何技术人员皆能够制作或使用本公开。对本公开的各种修改对本领域技术人员来说都将是显而易见的，且本文中所定义的普适原理可被应用到其他变体而不会脱离本公开的精神或范围。由此，本公开并非旨在被限定于本文中所描述的示例和设计，而是应被授予与本文中所公开的原理和新颖性特征相一致的最广范围。

Claims (22)

  1. 一种具备自主评价能力的人工智能系统，包括：
    评价能力模块，其配置成根据参考人群对于预定义事物的原始评价信息生成对于所述预定义事物的自主评价标准；
    所述评价能力模块还配置成根据所述自主评价标准对目标事物进行自主评价；以及
    决策模块，其配置成根据所述自主评价确定针对所述目标事物的输出。
  2. 如权利要求1所述的人工智能系统，其特征在于，还包括：
    感知模块，配置成获取所述参考人群对于预定义事物的原始评价信息；
    所述评价能力模块还配置成根据多个所述参考人群对于事物的原始评价信息分别生成相同的对于所述预定义事物的自主评价标准。
  3. 如权利要求1所述的人工智能系统，其特征在于，还包括：
    评价标准词汇模块，用于存储评价标准词汇信息；
    所述评价能力模块进一步配置成，将所述参考人群对于预定义事物的原始评价信息与所述评价标准词汇信息进行匹配，根据匹配结果生成对于所述预定义事物的所述自主评价标准。
  4. 如权利要求1所述的人工智能系统，其特征在于，还包括：
    评价标准词汇模块，用于存储评价标准词汇信息；
    所述评价能力模块还配置成，针对所述事物对用户进行提问；
    所述评价能力模块还配置成，将所述用户对于所述提问的回答与所述评价标准词汇信息进行匹配，根据匹配结果确定对于所述预定义事物的所述自主评价标准。
  5. 如权利要求1所述的人工智能系统，其特征在于，还包括：
    评价标准词汇模块，用于存储评价标准词汇信息；
    通信模块，所述评价能力模块还配置成经由所述通信模块从网络上获取对于预定义事物的网络评价；
    所述评价能力模块还配置成将所述网络评价与所述评价标准词汇信息进行匹配，根据匹配结果确定对于所述预定义事物的所述自主评价标准。
  6. 如权利要求3-5中任一项所述的人工智能系统，其特征在于，根据所述参考人群对于预定义事物的原始评价信息或所述用户对于所述提问的回答或将所述网络评价与所述评价标准词汇信息的匹配百分比确定所述自主评价标准。
  7. 如权利要求1所述的人工智能系统，其特征在于，所述评价能力模块还配置成生成对于已经为其生成自主评价标准的预定义事物的再次自主评价标准；
    所述评价能力模块还配置成将所述再次自主评价标准与现有自主评价标准加权相加以更新所述预定义事物的自主评价标准。
  8. 如权利要求7所述的人工智能系统，其特征在于，所述评价能力模块还配置成，将所述参考人群划分等级；
    所述加权相加还包括，根据所述参考人群的等级，赋予所述再次自主评价标准与原自主评价标准不同权重。
  9. 如权利要求1所述的人工智能系统，其特征在于，所述决策模块还配置成，将所述输出分为正向输出、负向输出；
    所述评价能力模块还配置成：
    将所述自主评价分为正向评价、低正向评价、负向评价、低负向评价；
    所述决策模块还配置成：
    若所述输出为所述正向输出，所述目标事物的自主评价为所述正向评价，或者，若所述输出为所述负向输出，所述目标事物的自主评价为所述负向评价，则加强所述输出的效果；
    若所述输出为所述正向输出，所述目标事物的自主评价为所述低负向评价，或者，若所述输出为所述负向动作，所述目标事物的自主评价为所 述弱正向评价，则减弱所述输出的效果；
    若所述输出为所述正向输出，所述目标事物的自主评价为所述负向评价，或者，若所述输出为所述负向输出，所述目标事物的自主评价为所述正向评价，则分别将所述输出更改为所述负向输出或正向输出。
  10. 一种具备自主评价能力的人工智能系统，其特征在于，包括：
    输出模块，其配置成执行对于目标事物的输出；
    评价能力模块，其配置成根据参考人群对于预定义事物的原始评价信息生成对于所述预定义事物的自主评价标准；
    所述评价能力模块还配置成根据所述自主评价标准对目标事物进行自主评价；以及
    决策模块，其配置成根据所述自主评价调整所述输出模块对于所述目标事物的所述输出。
  11. 如权利要求10所述的人工智能系统，其特征在于，所述决策模块还配置成，将所述输出分为正向输出、负向输出；
    所述评价能力模块还配置成：
    将所述自主评价分为正向评价、低正向评价、负向评价、低负向评价；
    所述根据所述自主评价调整所述输出模块对于所述目标事物的所述输出包括：
    若所述输出为所述正向输出，所述目标事物的自主评价为所述正向评价，或者，若所述输出为所述负向输出，所述目标事物的自主评价为所述负向评价，则加强所述输出的效果；
    若所述输出为所述正向输出，所述目标事物的自主评价为所述低负向评价，或者，若所述输出为所述负向动作，所述目标事物的自主评价为所述弱正向评价，则减弱所述输出的效果；
    若所述输出为所述正向输出，所述目标事物的自主评价为所述负向评价，或者，若所述输出为所述负向输出，所述目标事物的自主评价为所述正向评价，则分别将所述输出更改为所述负向输出或正向输出。
  12. 一种用于人工智能系统的自主评价方法，包括：
    根据参考人群对于预定义事物的原始评价信息生成对于所述预定义事物的自主评价标准；
    根据所述自主评价标准对目标事物进行自主评价；以及
    根据所述自主评价确定针对所述目标事物的输出。
  13. 如权利要求12所述的方法，其特征在于，还包括：
    获取所述参考人群对于预定义事物的原始评价信息；
    根据多个所述参考人群对于事物的原始评价信息分别生成相同的对于所述预定义事物的自主评价标准。
  14. 如权利要求12所述的方法，其特征在于，还包括：
    存储评价标准词汇信息；
    将所述参考人群对于预定义事物的原始评价信息与所述评价标准词汇信息进行匹配，根据匹配结果生成对于所述预定义事物的所述自主评价标准。
  15. 如权利要求12所述的方法啊，其特征在于，还包括：
    存储评价标准词汇信息；
    针对所述事物对用户进行提问；
    将所述用户对于所述提问的回答与所述评价标准词汇信息进行匹配，根据匹配结果确定对于所述预定义事物的所述自主评价标准。
  16. 如权利要求12所述的方法，其特征在于，还包括：
    存储评价标准词汇信息；
    经由所述通信模块从网络上获取对于预定义事物的网络评价；
    将所述网络评价与所述评价标准词汇信息进行匹配，根据匹配结果确定对于所述预定义事物的所述自主评价标准。
  17. 如权利要求14-16中任一项所述的方法，其特征在于，根据所述参考人群对于预定义事物的原始评价信息或所述用户对于所述提问的回答或将所 述网络评价与所述评价标准词汇信息的匹配百分比确定所述自主评价标准。
  18. 如权利要求12所述的方法，其特征在于，还包括：
    生成对于已经为其生成自主评价标准的预定义事物的再次自主评价标准；
    将所述再次自主评价标准与现有自主评价标准加权相加以更新所述预定义事物的自主评价标准。
  19. 如权利要求18所述的方法，其特征在于，还包括：
    将所述参考人群划分等级；
    所述加权相加还包括，根据所述参考人群的等级，赋予所述再次自主评价标准与原自主评价标准不同权重。
  20. 如权利要求12所述的方法，其特征在于，还包括：
    将所述输出分为正向输出、负向输出；
    将所述自主评价分为正向评价、低正向评价、负向评价、低负向评价；
    若所述输出为所述正向输出，所述目标事物的自主评价为所述正向评价，或者，若所述输出为所述负向输出，所述目标事物的自主评价为所述负向评价，则加强所述输出的效果；
    若所述输出为所述正向输出，所述目标事物的自主评价为所述低负向评价，或者，若所述输出为所述负向动作，所述目标事物的自主评价为所述弱正向评价，则减弱所述输出的效果；
    若所述输出为所述正向输出，所述目标事物的自主评价为所述负向评价，或者，若所述输出为所述负向输出，所述目标事物的自主评价为所述正向评价，则分别将所述输出更改为所述负向输出或正向输出。
  21. 一种用于人工智能系统的自主评价方法，其特征在于，包括：
    执行对于目标事物的输出；
    根据参考人群对于预定义事物的原始评价信息生成对于所述预定义事物的自主评价标准；
    根据所述自主评价标准对目标事物进行自主评价；以及
    根据所述自主评价调整所述输出模块对于所述目标事物的所述输出。
  22. 如权利要求21所述的方法，其特征在于，还包括：
    将所述输出分为正向输出、负向输出；
    将所述自主评价分为正向评价、低正向评价、负向评价、低负向评价；
    所述根据所述自主评价调整所述输出模块对于所述目标事物的所述输出包括：
    若所述输出为所述正向输出，所述目标事物的自主评价为所述正向评价，或者，若所述输出为所述负向输出，所述目标事物的自主评价为所述负向评价，则加强所述输出的效果；
    若所述输出为所述正向输出，所述目标事物的自主评价为所述低负向评价，或者，若所述输出为所述负向动作，所述目标事物的自主评价为所述弱正向评价，则减弱所述输出的效果；
    若所述输出为所述正向输出，所述目标事物的自主评价为所述负向评价，或者，若所述输出为所述负向输出，所述目标事物的自主评价为所述正向评价，则分别将所述输出更改为所述负向输出或正向输出。
PCT/CN2017/076267 2016-03-31 2017-03-10 具备评价能力的人工智能系统及其评价方法 WO2017166990A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610196845.1 2016-03-31
CN201610196845.1A CN107292466A (zh) 2016-03-31 2016-03-31 具备评价能力的人工智能系统及其评价方法

Publications (1)

Publication Number Publication Date
WO2017166990A1 true WO2017166990A1 (zh) 2017-10-05

Family

ID=59962579

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/076267 WO2017166990A1 (zh) 2016-03-31 2017-03-10 具备评价能力的人工智能系统及其评价方法

Country Status (2)

Country Link
CN (1) CN107292466A (zh)
WO (1) WO2017166990A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111680126A (zh) * 2020-06-10 2020-09-18 牛猛 一种基于人工智能的行为决策控制系统及方法
CN115407750A (zh) * 2022-08-12 2022-11-29 北京津发科技股份有限公司 人机协同智能系统决策能力的测评方法及系统
CN116739440A (zh) * 2023-08-16 2023-09-12 亚信科技（中国）有限公司 自智网络的评价方法、装置、电子设备及存储介质

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109484297A (zh) * 2019-01-08 2019-03-19 陶乐敏 一种人工智能车载安全检测系统

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101038581A (zh) * 2006-03-17 2007-09-19 国际商业机器公司 用于评价理解文件的难度的系统和方法
CN101859313A (zh) * 2009-04-08 2010-10-13 索尼公司 信息处理设备和方法及其程序
CN103632290A (zh) * 2013-12-02 2014-03-12 合肥工业大学 一种基于推荐概率融合的混合推荐方法
CN104699708A (zh) * 2013-12-09 2015-06-10 中国移动通信集团北京有限公司 一种客服机器人的自学习方法及装置
CN104699711A (zh) * 2013-12-09 2015-06-10 华为技术有限公司 一种推荐方法及服务器

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101038581A (zh) * 2006-03-17 2007-09-19 国际商业机器公司 用于评价理解文件的难度的系统和方法
CN101859313A (zh) * 2009-04-08 2010-10-13 索尼公司 信息处理设备和方法及其程序
CN103632290A (zh) * 2013-12-02 2014-03-12 合肥工业大学 一种基于推荐概率融合的混合推荐方法
CN104699708A (zh) * 2013-12-09 2015-06-10 中国移动通信集团北京有限公司 一种客服机器人的自学习方法及装置
CN104699711A (zh) * 2013-12-09 2015-06-10 华为技术有限公司 一种推荐方法及服务器

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111680126A (zh) * 2020-06-10 2020-09-18 牛猛 一种基于人工智能的行为决策控制系统及方法
CN111680126B (zh) * 2020-06-10 2023-06-20 牛猛 一种基于人工智能的行为决策控制系统及方法
CN115407750A (zh) * 2022-08-12 2022-11-29 北京津发科技股份有限公司 人机协同智能系统决策能力的测评方法及系统
CN115407750B (zh) * 2022-08-12 2023-11-21 北京津发科技股份有限公司 人机协同智能系统决策能力的测评方法及系统
CN116739440A (zh) * 2023-08-16 2023-09-12 亚信科技（中国）有限公司 自智网络的评价方法、装置、电子设备及存储介质
CN116739440B (zh) * 2023-08-16 2023-11-17 亚信科技（中国）有限公司 自智网络的评价方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
CN107292466A (zh) 2017-10-24

Similar Documents

Publication Publication Date Title
WO2017166990A1 (zh) 具备评价能力的人工智能系统及其评价方法
US10049307B2 (en) Visual object recognition
CN105913121B (zh) 神经网络训练方法和设备以及识别方法和设备
Han et al. Strength modelling for real-worldautomatic continuous affect recognition from audiovisual signals
US9805493B2 (en) Social identity models for automated entity interactions
US11880754B2 (en) Electronic apparatus and control method thereof
Srivastava et al. UAV surveillance for violence detection and individual identification
KR20200010680A (ko) 기계학습 모델을 이용하여 n개의 프레임에 기초하여 대상의 감정인식을 수행하는 감정인식 시스템, 방법, 및 컴퓨터-판독가능매체
WO2018210323A1 (zh) 一种用于提供社交对象的方法与设备
KR20200115706A (ko) 딥러닝 기반의 이미지 영역화 방법 및 그 장치
Alameda-Pineda et al. Multimodal behavior analysis in the wild: An introduction
US11688158B1 (en) Method and system for unique, procedurally generated digital objects via few-shot model
CN113408693A (zh) 识别图像的方法和设备
Jaroli et al. Deep learning model based novel semantic analysis
JP2023536802A (ja) 顔認識のデータセットアウェア及び不変学習
US20200057758A1 (en) Server and control method thereof
US20230360027A1 (en) Method and system for unique, procedurally generated extended reality environment via few-shot model
JP2023527341A (ja) プロトタイプオプションの発見による解釈可能な模倣学習
Roy et al. Facial Recognition Techniques and Their Applicability to Student Concentration Assessment: A Survey
Lee et al. Combining voice and image recognition for smart home security system
Wang Improved Generative Adversarial Networks for Student Classroom Facial Expression Recognition
US20240312360A1 (en) Apparatus and method for an education platform and ecosystem using extended reality
US11949967B1 (en) Automatic connotation for audio and visual content using IOT sensors
US20240311915A1 (en) Systems and methods for incentivizing behavior using data projection
US20240236091A1 (en) Systems and methods for recognizing new devices

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

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

Ref document number: 17773021

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 17773021

Country of ref document: EP

Kind code of ref document: A1