CN109191293B - Artificial intelligence service system and method based on intelligent contract and general evidence - Google Patents
Artificial intelligence service system and method based on intelligent contract and general evidence Download PDFInfo
- Publication number
- CN109191293B CN109191293B CN201810901731.1A CN201810901731A CN109191293B CN 109191293 B CN109191293 B CN 109191293B CN 201810901731 A CN201810901731 A CN 201810901731A CN 109191293 B CN109191293 B CN 109191293B
- Authority
- CN
- China
- Prior art keywords
- expert
- rules
- data
- agreed
- new
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 121
- 238000000034 method Methods 0.000 title claims abstract description 54
- 238000011156 evaluation Methods 0.000 claims abstract description 37
- 230000007246 mechanism Effects 0.000 claims abstract description 14
- 238000012795 verification Methods 0.000 claims description 45
- 238000001514 detection method Methods 0.000 claims description 12
- 238000002372 labelling Methods 0.000 claims description 11
- 238000012360 testing method Methods 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 9
- 230000008901 benefit Effects 0.000 claims description 8
- 230000006870 function Effects 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 abstract description 24
- 238000005516 engineering process Methods 0.000 abstract description 5
- 201000010099 disease Diseases 0.000 description 9
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 9
- 208000024891 symptom Diseases 0.000 description 8
- 238000012549 training Methods 0.000 description 8
- 230000015654 memory Effects 0.000 description 7
- 230000006399 behavior Effects 0.000 description 5
- 238000013499 data model Methods 0.000 description 5
- 230000036760 body temperature Effects 0.000 description 4
- 238000011161 development Methods 0.000 description 4
- 210000004072 lung Anatomy 0.000 description 4
- 230000003211 malignant effect Effects 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 208000004998 Abdominal Pain Diseases 0.000 description 3
- 206010063659 Aversion Diseases 0.000 description 3
- 206010011224 Cough Diseases 0.000 description 3
- 206010056342 Pulmonary mass Diseases 0.000 description 3
- 206010037660 Pyrexia Diseases 0.000 description 3
- 238000012550 audit Methods 0.000 description 3
- 230000002146 bilateral effect Effects 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 3
- 210000002741 palatine tonsil Anatomy 0.000 description 3
- 239000007787 solid Substances 0.000 description 3
- 239000013598 vector Substances 0.000 description 3
- 206010002383 Angina Pectoris Diseases 0.000 description 2
- 102000001554 Hemoglobins Human genes 0.000 description 2
- 108010054147 Hemoglobins Proteins 0.000 description 2
- 201000008235 Mycoplasma pneumoniae pneumonia Diseases 0.000 description 2
- 208000037656 Respiratory Sounds Diseases 0.000 description 2
- 206010057190 Respiratory tract infections Diseases 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 239000002221 antipyretic Substances 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 210000003651 basophil Anatomy 0.000 description 2
- 238000001816 cooling Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000013480 data collection Methods 0.000 description 2
- 238000013524 data verification Methods 0.000 description 2
- 206010012601 diabetes mellitus Diseases 0.000 description 2
- 210000003979 eosinophil Anatomy 0.000 description 2
- 210000000265 leukocyte Anatomy 0.000 description 2
- 210000004698 lymphocyte Anatomy 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 210000000440 neutrophil Anatomy 0.000 description 2
- 238000003909 pattern recognition Methods 0.000 description 2
- 210000003800 pharynx Anatomy 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 108090000623 proteins and genes Proteins 0.000 description 2
- 238000012552 review Methods 0.000 description 2
- 238000000638 solvent extraction Methods 0.000 description 2
- 230000008961 swelling Effects 0.000 description 2
- 208000035473 Communicable disease Diseases 0.000 description 1
- 206010067715 Gastrointestinal sounds abnormal Diseases 0.000 description 1
- 208000035211 Heart Murmurs Diseases 0.000 description 1
- 208000001572 Mycoplasma Pneumonia Diseases 0.000 description 1
- 241000202934 Mycoplasma pneumoniae Species 0.000 description 1
- 206010028980 Neoplasm Diseases 0.000 description 1
- 206010035724 Pneumonia mycoplasmal Diseases 0.000 description 1
- 206010046306 Upper respiratory tract infection Diseases 0.000 description 1
- 206010047700 Vomiting Diseases 0.000 description 1
- 230000000172 allergic effect Effects 0.000 description 1
- 208000010668 atopic eczema Diseases 0.000 description 1
- 238000002555 auscultation Methods 0.000 description 1
- 210000004369 blood Anatomy 0.000 description 1
- 239000008280 blood Substances 0.000 description 1
- 230000036772 blood pressure Effects 0.000 description 1
- 206010006451 bronchitis Diseases 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004138 cluster model Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000035622 drinking Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000005337 ground glass Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 208000015181 infectious disease Diseases 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 230000029058 respiratory gaseous exchange Effects 0.000 description 1
- 208000020029 respiratory tract infectious disease Diseases 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000013526 transfer learning Methods 0.000 description 1
- 210000002700 urine Anatomy 0.000 description 1
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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
-
- 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
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/08—Payment architectures
- G06Q20/14—Payment architectures specially adapted for billing systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/018—Certifying business or products
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Engineering & Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- Development Economics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Technology Law (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an artificial intelligence service method and system based on intelligent contracts and passes. The artificial intelligence service method comprises the following steps: uploading expert agreed rules or data driven models to an evaluation system and verifying; binding the authenticated expert agreed rules or the evaluation systems corresponding to the data driving models to intelligent contracts, and writing block chains to issue to an artificial intelligent service system; when input data call artificial intelligent service to predict, expert negotiation rules or data driving models applicable to an artificial intelligent service system call mode library obtain a prediction result; and paying the pass to the artificial intelligence service system according to the intelligent contract after the user selects the prediction result. The method applies the blockchain technology to the professional field, realizes the sharing of professional data and expert experience, and effectively ensures the legal rights and interests of industry experts, industry data contributors, algorithm personnel and system maintainers by a certification mechanism.
Description
Technical Field
The invention relates to an artificial intelligence service system based on intelligent contracts and certificates (Token), and simultaneously relates to an artificial intelligence service method based on intelligent contracts and certificates, belonging to the technical field of artificial intelligence.
Background
Currently, AI-driven services are actively developed in various fields such as medical, financial, unmanned, electronic commerce, etc. AI (Artificial Intelligence ) is generally divided into two main branches, rule-based and data-driven. In the data driving method, classical statistical pattern recognition supported by Bayes theory of probability, clustering of space vector similarity and the like and with strong interpretation, and deep learning networks such as CNN, LSTM and the like with obvious black box characteristics based on generalized neural network (algorithm of minimizing energy function including Boltzmann machine and the like) development are also included.
Currently, many applications often merge several classes as well. For example, knowledge-graphs appear primarily as rule-based models, but many knowledge-graph rules or state transfer functions combine interpretable probability and cluster models.
Whether rule-based methods, interpretable probabilistic models (Bayesian Net, etc.), or models that rely more on deep learning, reinforcement learning, and transfer learning of a large number of data driven discovery patterns, require a closed loop in practice, evolving constantly. However, in the field of artificial intelligence in the professional fields such as medical treatment, data and knowledge are not applied and shared. A great deal of knowledge possessed by some experts cannot be widely applied, and localization of data and privatization of knowledge lead to uneven technical levels in different regions. The expert does not have the power to consume time and effort to share the experience accumulated for years on a large scale, because no corresponding return is obtained.
Therefore, a method for ensuring the safe sharing of knowledge and experience and ensuring that contributors obtain new AI-enabled professional domain services with corresponding rewards is an urgent need.
Token, as a digital form of proof of interest, may be generated and circulated in a centralized system, or may be generated and circulated in a decentralized system, such as a blockchain (public chain or alliance chain). Token, in combination with blockchain, can determine authenticity and asset uniqueness through encryption algorithms (authenticity, tamper resistance, privacy protection) and distributed account books, and circulate through consensus algorithms.
The blockchain technology is a novel decentralization protocol, related data can be safely stored, information cannot be forged and tampered, the historical retrospection is strong, transaction and information data access authentication on the blockchain are completed by all nodes on the blockchain together, and consistency is guaranteed by a continuously innovative consensus algorithm. A common ledger (in-chain) can be maintained on the blockchain for storing data of all users on the blockchain, and the common ledger is visible to any node on the storage block, thereby preventing tampering of the data. However, currently, blockchain techniques are rarely applied to AI-enabled professional field services.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide an artificial intelligent service system based on intelligent contracts and general evidence.
Another technical problem to be solved by the invention is to provide an artificial intelligence service method based on intelligent contracts and passes.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
according to a first aspect of an embodiment of the present invention, there is provided an artificial intelligence service system based on intelligent contracts and certificates, including a knowledge uploading module, a knowledge verifying module, a publishing module, a predicting module and a payment module;
the knowledge uploading module is used for uploading expert agreed rules or a data driving model to the evaluation system;
the knowledge verification module is used for verifying the knowledge transmitted by the knowledge uploading module and transmitting expert agreed rules or data driven models passing verification to the issuing module;
the issuing module is used for binding the authenticated expert agreed rules or the evaluation systems corresponding to the data driving models to intelligent contracts, writing block chains and issuing to the artificial intelligent service system;
when user input data call artificial intelligent service to predict, the prediction module predicts according to expert agreed rules or data driving models issued by the issuing module and outputs a prediction result to the user;
And the payment module pays the pass to the artificial intelligence service system from the user account according to the intelligent contract according to the prediction result of the prediction module selected by the user, and performs transaction information release through the release module.
Wherein preferably, the artificial intelligence service system based on intelligent contract and pass further comprises a conflict detection module;
the conflict detection module is used for carrying out conflict detection on the uploaded expert agreed rules or the data driving model and judging whether the uploaded expert agreed rules or the uploaded data driving model are new expert agreed rules or new data driving models.
According to a second aspect of an embodiment of the present invention, there is provided an artificial intelligence service method based on an intelligent contract and a certification, including the steps of:
uploading expert agreed rules or data driven models to an evaluation system, and performing evaluation verification;
storing expert agreed rules or data driven models which pass the evaluation and verification as verified modes in a mode library, and issuing the modes to a corresponding AI service system; simultaneously binding a benefit distribution mechanism to an intelligent contract, and writing the benefit distribution mechanism into a blockchain;
when input data call AI service to predict, the AI service system calls expert negotiation rules or data driving models which are applicable to the mode library to obtain a prediction result and outputs the prediction result;
After the user selects a rule or model and its corresponding prediction, the input, output and selected behavior records of the entire data are counted into the blockchain and Token is paid to each contributor to the model according to the smart contract.
Wherein preferably, after the expert agrees on the rule or the data-driven model is uploaded to the evaluation system, the method further comprises the following steps before the verification:
and carrying out conflict detection on the uploaded expert agreed rules or data driven models, and judging whether the expert agreed rules or the data driven models are newly added and valid.
Preferably, the conflict test is performed on the uploaded expert negotiation rules to determine whether the expert negotiation rules are newly added valid expert negotiation rules, and the method comprises the following steps:
s101, judging whether an old expert agreed rule which is repeated or partially repeated with the uploaded new expert agreed rule exists in the AI service system, and if so, turning to the step S102; otherwise, turning to step S103; wherein, repeated or partial repeated refers to that the applicable limiting conditions of the input data are the same or partially overlapped;
s102, comparing with the old expert agreed rules, judging whether the new expert agreed rules are updated, if so, turning to a step S103; otherwise, refusing the uploading of the new expert agreed rules;
S103, expert evaluation verification is carried out on the uploaded new expert negotiation rules.
Wherein it is preferably determined whether the expert negotiation rules update one or more of the input comprising the new expert negotiation rules, whether the new expert negotiation rules add new data objects, whether the new expert negotiation rules further subdivide the input old data objects, whether the new expert negotiation rules add attributes to the input objects, whether the new expert negotiation rules modify attributes, whether the new expert negotiation rules further divide the old expert negotiation rules, and whether a condition or parameter in the rules changes.
Wherein preferably, the step of evaluating and verifying the uploaded expert agreed rule comprises the following steps:
determining expert committee members for auditing according to the field to which the uploaded expert negotiation rules belong;
and setting a verification threshold, and when the number of expert committee members qualified by the expert negotiation rules uploaded by verification is greater than the verification threshold, evaluating that the expert negotiation rules uploaded by the verification are qualified, and evaluating that the verification is passed.
Preferably, the conflict test is performed on the data driving model uploaded by the algorithm team to judge whether the model is a newly added data driving model, and the method comprises the following steps:
s111, judging diseases of the labeling data sets continuously accumulated by the system by the original data driving model and the new data driving model at the same time, and labeling the consistency of the new data driving model and each old model one by one;
S112, rejecting the release of the new data driving model when an existing old model exists and the output of the old model is completely consistent with the output of the new model; otherwise, go to step S113;
s113, evaluating and verifying the uploaded new data driving model.
Preferably, the new data driving model comprises one or a combination of a new data driving model obtained by model optimization training based on the old annotation data set existing in the system and a new data driving model obtained by new annotation data training in the system.
Preferably, the new data driving model is evaluated and verified, and the method comprises the following steps:
multiplying the false alarm rate and the false alarm rate by corresponding cost functions to obtain a final loss expected value, and if the final loss expected value of the new data driving model is smaller than the final loss expected value of the original data driving model and is lower than a model prediction loss function threshold preset by the system, evaluating and verifying.
Preferably, the newly added expert agreed rules or data driven models which pass the evaluation verification are stored in a mode library as verified modes and are released to the corresponding AI service systems; and simultaneously, the benefit distribution mechanism is bound with the intelligent contract and written into the blockchain.
Preferably, when the input data call the AI service for prediction, an expert agreed rule or a data driving model applicable in the AI service system call mode library obtains a prediction result and outputs the prediction result, and the method comprises the following steps:
when the input data call the AI service for prediction, the AI service system searches a corresponding expert negotiation rule and a data driving model according to the input data;
judging the input labeling data according to the found expert negotiation rules and the data driving model to obtain a prediction result;
and outputting and displaying the selected mode (expert agreed rules and data driven models) and the prediction result.
Preferably, when the selected prediction result corresponds to a plurality of expert agreed rules or data driven models, the AI service system matches the expert agreed rules or data driven models selected for the lowest transaction price.
Wherein preferably, the artificial intelligence service method further comprises the following steps:
and when the prediction result is judged to be wrong, feeding back to the AI service system. The AI service system will write to the blockchain and to the annotation database.
Preferably, when the number of times of the predicted result errors of the user feedback system exceeds a specific number threshold, the predicted result data sets are checked to determine whether the corresponding expert agreed rule or the data driving model has errors on the data examples fed back by the users or the users themselves determine the errors.
Wherein preferably, when the user feedback system predicts that the result is incorrectly fed back, a Token reward is given to the user based on the smart contract.
Wherein preferably, the artificial intelligence service method further comprises the following steps:
token is assigned to different types of contributors according to smart contracts and pre-established expert agreed rules and recorded into the accounting nodes of the blockchain.
According to the artificial intelligence service system and the method provided by the invention, the blockchain technology is applied to the professional field, so that the sharing of professional data and expert experience is realized, and the legal rights and interests of industry experts, industry data contributors, algorithm personnel and system maintainers are effectively ensured by a certification mechanism.
Drawings
FIG. 1 is a schematic diagram of an artificial intelligence service system based on intelligent contracts and passes provided by the invention;
FIG. 2 is a flow chart of an artificial intelligence service method based on intelligent contracts and passes provided by the invention.
Detailed Description
The technical contents of the present invention will be described in detail with reference to the accompanying drawings and specific examples.
The invention firstly provides an artificial intelligence service system based on intelligent contracts and passes. As shown in fig. 1, the system at least comprises a knowledge uploading module, a knowledge verifying module, a publishing module, a predicting module and a payment module.
The knowledge uploading module is used for uploading expert negotiation rules or a data driving model (knowledge) to the evaluation system; the expert agreements rules are that the expert uploads expert agreements rules in the candidate rule base to an evaluation system for evaluation by an expert committee and verification is carried out; the AI service system can be issued after the verification is passed; the data driving model is obtained by carrying out algorithm modeling by an algorithm team according to the historical operation data of the AI service system and the data uploaded by the expert, and the data driving model is uploaded to an evaluation system for evaluation verification, and is released to the AI service system after the verification is passed. The detailed verification process is described in detail later.
The knowledge verification module is used for verifying the knowledge transmitted by the knowledge uploading module and transmitting expert agreed rules or data driven models passing verification to the issuing module; the issuing module is used for binding the authenticated expert agreed rules or the evaluation systems corresponding to the data driving models to the intelligent contracts, and writing the intelligent contracts into the block chain to issue the intelligent contracts to the AI service system.
When the user inputs data to call the AI service for prediction, the prediction module predicts according to expert agreed rules or a data driving model issued by the issuing module, and outputs a prediction result to the user; the payment module pays Token from the user account to the AI service system according to the intelligent contract according to the prediction result of the prediction module selected by the user; and sending the transaction information to the release module for releasing the transaction information.
In the embodiment provided by the invention, the artificial intelligence service system based on the intelligent contract and the evidence further comprises a conflict detection module. When the AI service system is established, an expert is required to upload expert agreed rules or data driven models to the blockchain evaluation system. In the embodiment provided by the invention, after the knowledge uploading module uploads the expert agreed rule or the data driving model to the evaluation system, the conflict detection module is required to carry out conflict detection on the uploaded expert agreed rule or data driving model before the knowledge verification module verifies the uploaded expert agreed rule or data driving model, so as to judge whether the expert agreed rule or the data driving model is a new expert agreed rule or a new data driving model. The expert agreed rules or data driven models are guaranteed not to conflict with existing expert agreed rules or data driven models in the AI service system. I.e. the same expert agreed rules or data driven models are not present in the AI service system.
As shown in fig. 2, the artificial intelligence service method based on intelligent contract and pass provided by the invention comprises the following steps: firstly, uploading expert agreed rules or a data driving model to an evaluation system, and verifying; binding the authenticated expert agreed rules or the evaluation systems corresponding to the data driving models to intelligent contracts, writing block chains and issuing to an AI service system; then, when the user inputs data to call the AI service for prediction, an expert in an AI service system call mode library agrees on rules or a data driving model to obtain a prediction result, and the prediction result is output to the user; after the user selects the prediction result, paying Token to the AI service system according to the intelligent contract; finally, the transaction information is written into the blockchain and the blockchain is sent to the AI service system. This process is described in detail below.
S1, uploading expert agreed rules or data driven models to an evaluation system, and verifying.
When the AI service system is established, an expert is required to upload expert agreed rules or data driven models to the blockchain evaluation system. In the embodiment provided by the invention, after the expert agrees to upload the rule or the data driving model to the evaluation system, the method further comprises the following steps:
and carrying out conflict detection on the uploaded expert agreed rules or data driven models, and judging whether the expert agreed rules or the data driven models are new expert agreed rules or new data driven models.
In the embodiment provided by the invention, after the expert agreed rule or the data driven model is uploaded to the expert agreed rule evaluation system, before the expert agreed rule or the data driven model is verified, the loaded expert agreed rule or the data driven model is subjected to conflict verification, so that the expert agreed rule or the data driven model is ensured not to conflict with the existing expert agreed rule or data driven model in the AI service system. I.e. the same expert agreed rules or data driven models are not present in the AI service system.
The method comprises the following steps of:
S101, judging whether an old expert agreed rule which is repeated or partially repeated with the uploaded new expert agreed rule exists in the AI service system, and if so, turning to the step S102; otherwise, turning to step S103; wherein repeated or partially repeated refers to the same or partially overlapping constraints of the applicable input data.
S102, comparing with the old expert agreed rules, judging whether the new expert agreed rules are updated, if so, turning to a step S103; otherwise, the new expert's agreed rules are refused to be uploaded.
In the embodiment provided by the invention, compared with the old expert negotiation rule, whether the new expert negotiation rule is updated is judged, wherein the method comprises one or more of adding a new object to the new expert negotiation rule, further subdividing the old object by the new expert negotiation rule, adding an attribute to the object of the new expert negotiation rule, modifying the attribute of the new expert negotiation rule, further dividing the old expert negotiation rule by the new expert negotiation rule, and changing a certain condition or parameter in the rule.
Specifically, compared with the old expert agreed rule, judging whether the new expert agreed rule is updated or not includes:
a) Adding new objects or further subdividing old objects, or adding attributes or modifying attributes to objects. For example: the invention of gene detection technology not only adds new objects such as various description types of genes, but also carries out finer typing on a plurality of tumor diseases.
CT examination reveals nodules that are further classified as solid nodules, ground glass shadows, and semi-solid nodules. Further, the properties of the nodule include various image histology features such as HU mean, standard deviation, lowest HU value, highest HU value, kurtosis, edge sharpness, edge blurriness, etc.
Diabetes mellitus: can be divided into several types of diabetes and can be further subdivided.
b) New expert negotiation rules are added or revised. The new expert's agreed rules are revised, typically by further subdividing the old expert's agreed rules, making it more accurate for narrower situations.
For example, in compliance with patient population descriptions, symptom combination descriptions (A1, A2, … … A8); the states described by the signs (B1, B2, B3) correspond to the diseases D1, D2, D3, originally with probabilities P1, P2, P3, respectively. Later replaced with more accurate (A1, A2, … … A8); signs (B1, B2, B3) and biochemical physical examination items (C1, C2, C3, C4) corresponding to disease D1; and (A1, A2, … … A8), signs (B1, B2, B3, B5) and biochemical physical examination items (C1, C2, C3, C4), corresponding to disease D2. Examples are as follows:
child patients, women, 11 years old, han nationality, frequent XX village, good constitution, no repeated history of respiratory tract infection, no allergic history, and no infectious disease contact history.
Reasoning based on old rules: symptoms (fever, aversion to cold, cough); sign (body temperature 39 ℃, 32 times/min breath, reddish throat, no obvious swelling of bilateral tonsils) -upper respiratory infection, acute bronchitis, mycoplasmal pneumonia … …
After the new expert agreements rules are added or corrected and the old expert agreements rules are further subdivided, various indexes of physical sign-dry-wet royalty and blood routine and urine routine in biochemical inspection and inspection are added, so that the method is more accurate and suitable for more refined conditions.
One of the inference paths based on update rules: symptoms (fever, aversion to cold, cough: lighter); sign (body temperature 39 ℃ C. -antipyretic cooling effect, 32 times/min breath-no smell and dry and wet royalty, slightly red pharynx, no obvious enlargement of bilateral tonsils), test index (white blood cells 4.4x109/L, hemoglobin 128g/L, total number of platelets 191x109/L, neutrophil percentage 63%, lymphocyte percentage 26.6%, eosinophil percentage 1.10%, basophil percentage 0.0%) ".
One of the inference paths based on update rules: symptoms (fever: intermittent-1 week, aversion to cold, cough: paroxysmal-irritative); sign (body temperature 39 ℃ C. -antipyretic cooling has no obvious change, respiratory sound of the right lung is reduced, respiratory sound of the left lung is clear, dry and wet roar sound is not smelled, pharyngeal portion is reddish, bilateral tonsils have no obvious swelling), inspection index (white blood cell 4.4x109/L, hemoglobin 128g/L, total platelet count 191x109/L, neutrophil percentage 63%, lymphocyte percentage 26.6%, eosinophil proportion 1.10%, basophil proportion 0.0%, mycoplasma pneumoniae IgM antibody (+)), -mycoplasma pneumonia
Description:
patient population descriptions are age (stage), gender, ethnicity, general location, disease location, medical history, etc.
Single symptom description example: abdominal pain-site: periumbilical pain; traits: angina pectoris; frequency of: an array hair style; the inducement is as follows: after drinking.
The symptom combination is a combination described in detail as a single symptom above, for example (abdominal pain: periumbilical-paroxysmal-angina; emesis: gruel-postmeal-night) is a combination of two symptoms.
Signs include body temperature, pulse, respiration, blood pressure, and also include auscultation to obtain various sound signs (heart murmur, borborygmus, etc.), responses to tactile pressure, etc.
And S103, performing expert verification on the uploaded new expert negotiation rule. Preferably, when the expert committee member carries out auditing modification on the expert agreed rule, a new expert agreed sub-branch rule is established by modifying the uploaded expert agreed rule, the sub-branch rule becomes a new verified rule after verification, and the expert for establishing the sub-branch rule is the creator of the new evaluation system.
And after the uploaded expert agreed rules or data driven models are subjected to conflict verification, the expert agreed rules or data driven models are verified. The uploaded expert negotiation rules can be established for machines to analyze professional lessons through NLP (Natural Language Processing ) or various guidance expert negotiation rules, and can also be established for experts based on own knowledge and experience. After the expert agreed rules are uploaded, the expert committee members in the system are required to be subjected to auditing and correction through an AI service system according to the automatic assignment of the field to which the expert agreed rules belong, and after verification is passed, the intelligent contracts set by the blockchain can be bound and then issued. In the embodiment provided by the invention, the expert committee member in the system verifies the uploaded expert negotiation rule, and specifically comprises the following steps:
Determining expert committee members for auditing according to the field to which the uploaded expert negotiation rules belong; in the embodiments provided by the present invention, a partial number of experts within the industry are required to conduct the auditing together.
Setting a verification threshold, and when the number of members agreeing to the rule is larger than the verification threshold in the expert committee members qualified by the uploaded expert agreed rule, qualifying the uploaded expert agreed rule;
otherwise, the expert agreed rules are refused to be released to the AI service system.
In the embodiment provided by the invention, when the expert committee member examines the expert agreed rules uploaded by the expert, the sub-branch rules are established by correcting the uploaded expert agreed rules, and then the established sub-branch expert agreed rules are equivalent to the newly established expert agreed rules, and the other expert committee members are required to verify. The verification process of the branch sub-rule agreed by the expert is the same as the verification process of the expert agreed rule uploaded above, and will not be described again here. After verification, the branch sub-rule has a new evaluation system, and an expert establishing the branch sub-rule is established as the new rule owner.
The method comprises the following steps of:
s111, performing disease judgment on the labeling data set continuously accumulated by the system by the original data driving model and the new data driving model to obtain a false alarm rate;
the new data driven model is different from the new data. In general, adding new data (valid), i.e., adding some data that the model predicts unparallel, and adding some labeling attributes (labels), will bring a new data driven model. However, even without the newly added data, there is the possibility of further model optimization by the algorithm personnel. This newly added, better model also needs to be accepted by the audit.
In the embodiment provided by the invention, the new data driving model can be obtained by model optimization and training based on the old marked data set existing in the system; the model can be a new data driving model which is obtained by training based on new data in the system.
In the embodiment provided by the invention, the newly added data comprises one or two of newly added labels of the existing data samples and data which does not accord with the prediction of the original data driving model (the existing data driving model of the AI service system) actually occurs.
The case of adding a label to an existing data sample is as follows: the lung nodule labels of a large number of chest lung CT belong to the new label of the existing data sample if the benign and malignant grades and the nodule types are further labeled.
In the newly added data, data which does not coincide with the prediction of the old data-driven model actually occurs. Examples are: and (3) marking lung nodules of a large number of chest lung CT, automatically processing an input image after a model is established, identifying the nodules, and predicting the benign and malignant grades and the types of the nodules. If a node benign or malignant judgment, which is commonly recognized by the physician's cross judgment, is encountered, and the model prediction is inconsistent, then these labeled samples need to be of additional interest for further analysis as samples to provide evolutionary improvement of the model.
When a new data driving model is obtained based on the training of the new data in the system, AI algorithm personnel train the new data driving model according to a marked sample data set containing the new data; the AI algorithm personnel trains a new data driving model process according to the marked sample data set containing the new added data, and the process of verification is similar to that of the subsequent uploading of the data driving model to the evaluation system, namely, the AI algorithm personnel establishes various models and performs a closed test according to the marked sample data set containing the new added data; setting a model precision threshold, and obtaining an updated data-driven model when the test accuracy of the established model reaches the model precision threshold, which will not be described in detail.
S112, the processor judges whether an existing old model exists, and the output of the old model is completely consistent with the output of the new model. If so, release of the new data driven model is denied. If no such model exists, the process goes to step S113.
And S113, the processor evaluates and verifies the uploaded new data driving model. The AI service system prohibits the plagiarism specialist from negotiating rules and data driven models. All experts that have proven to have no difference in input and output results agree on rules or data driven models that will automatically mask the following creator according to the first come first go rule. But the same expert agrees on rules and data driven models, if the new data driven model is realized faster, the system automatically gives the updater a certain division to pay according to Token based on the expert agrees on rules in the division mechanism.
In the embodiments provided herein, an expert agrees that rule founders or data driven model founders enjoy the "patent-like" treatment within the blockchain, prohibiting any plagiarism and imitation for a period of time. Beyond this time period, any newly created expert agrees on rules or data driven models, which can be modeled and run at low cost.
When the data driven model is uploaded to the assessment system, verification is performed, comprising the steps of:
The evaluation system multiplies the misreport rate and the missing report rate of the data driving model by a corresponding cost function to obtain a final loss expected value, and if the final loss expected value of the new data driving model is smaller than the final loss expected value of the original data driving model and is lower than a model prediction loss function threshold value preset by the system, the new data driving model is stored in a mode library as a verified mode and is released to a corresponding AI service system; and simultaneously, the benefit distribution mechanism is bound with the intelligent contract and written into the blockchain.
False positives and false negatives of the data driven model are based on sample data sets in the annotation database.
In general, an industry expert uploads a labeled sample dataset to an evaluation system; if a data driven model needs to be constructed, a large amount of historical data is needed for analysis and processing. The training data is marked by an expert, namely a batch of marked sample data sets approved by industry experts are obtained.
The AI algorithm personnel establishes a model and adjusts parameters based on the labeling data, and performs a closed test according to the labeled sample data set; the model established by AI algorithm personnel can be a model established by classical statistical pattern recognition with strong interpretation supported by theory such as Bayes theory of probability and clustering of space vector similarity, and also can be a model established by a deep learning network with obvious black box characteristics based on generalized neural network development. There is no limitation on the method of constructing the model.
For example: the expert-labeled CT medical image sample, namely, the doctor labels the space position of a focus (such as a lung nodule) in the three-dimensional CT image in the image, and the focus type and focus attribute (benign and malignant grade and the like) of the position. And training various models by AI algorithm personnel according to the marked CT medical image samples, and passing verification when the test accuracy of the established model reaches a model accuracy threshold.
S2, storing the mode-expert negotiation rule or the data driving model which passes the verification in a mode library and issuing the mode-expert negotiation rule or the data driving model to a corresponding AI service system; and simultaneously, the benefit distribution mechanism is bound with the intelligent contract and written into the blockchain.
In the embodiment provided by the invention, the transaction price of the intelligent contract is divided into the determination according to the contribution degree to the AI service system. The creator (including the updater) of the expert agreed rule becomes the owner of the evaluation system and will have the virtual red out right of the expert agreed rule. I.e. whenever the expert agreed rule is invoked, the AI service system will pay a certain proportion or a certain number of Token to the creator of the expert agreed rule. The other Token can be used as tax paid by the evaluation system to the whole AI service system for compensating the auditing and correction of the rule and the operation, maintenance and popularization of the AI service system.
With respect to the pricing of each expert-agreed rule invocation, the AI service system will give the expert-agreed rule's virtual equity owner a market guidance price. How priced is in particular is decided by the expert's agreed rule holder. Market guidance prices, calling pricing of other similar expert agreed rules corresponding to the expert agreed rules in the AI service system, and labor cost for judging the corresponding expert agreed rules corresponding to the outside of the AI service system.
With respect to the creation of data driven models, a common effort is required by both the original labeling dataset and the AI algorithm personnel. Thus, one data driven model creator is two groups of people. One party is a collection of data contributors, occupying a share. The individual virtual red rights of the share are calculated according to the contribution proportion (namely, the proportion of the weighted contribution value of each person contribution and audit data to the weighted contribution total value of the whole annotation sample data). The other party contributes to the algorithm of the algorithm personnel. The data driven model call pricing is similar to expert agreed rule call pricing and will not be described in detail here.
And S3, when the user inputs data to call the AI service for prediction, an expert agreed rule or a data driving model which is applicable to the AI service system call mode library obtains a prediction result and outputs the prediction result to the user.
All expert agreed rules or data driven models issued by the AI service system will be integrated into the service. When a user inputs certain types of data (images, characters, time sequence data, vectors and the like) to an AI service system to call AI service for prediction and identification, the AI service system integrates various expert negotiation rules or data driving models (the system judgment is possibly correct) to obtain a prediction result, and the prediction result is output to the user, and the method specifically comprises the following steps:
when the user inputs data to call the AI service for prediction, the AI service system searches corresponding expert negotiation rules and data driving models according to the input data.
Judging the input labeling data according to the found expert negotiation rules and the data driving model to obtain a prediction result; in the embodiment provided by the invention, the prediction result is represented as a statistics calculation of a false alarm rate or a final loss expected value. The statistics are used for judging the diseases of the marked data sets continuously accumulated by the system through expert agreed rules or a new data driving model, wherein each data contains a final mark, and the false alarm rate and the expected loss value of the diseases are obtained.
And outputting and displaying the prediction result selected by the user. Wherein, the show includes: token equity price, disease misinformation rate, loss expectation value and the like of the prediction result are displayed, and the output data of the dimensions can be ranked.
In the embodiment provided by the invention, when a user selects to use a certain predicted result, the same output result generated by different data-driven models or expert negotiation rules (namely, different data-driven models or expert negotiation rules can give the same result on certain specific input data) is possible, and the AI service system automatically selects the expert negotiation rule or the data-driven model with the lowest price to match, and gives the corresponding Token incentive to the evaluation system corresponding to the data-driven model.
S4, after the user selects the prediction result, the Token is automatically paid to the AI service system. Only after the user (AI service caller) pays the equivalent Token can the prediction result and details be consulted.
If the user selects one of the pushed multiple prediction results, the corresponding Token is paid to the AI service system according to the intelligent contract set on the blockchain according to the expert negotiation rule or the price of the data driving model corresponding to the prediction result.
In the process that the user uses the whole industry AI service, a conclusion that a certain expert agrees on a rule or a certain data driving sub-model is adopted, the paid Token corresponding to the service is distributed by a system, and a specific Token is distributed to various types of contributors of the rule or the model of the node; the remaining Token, the system reserves the expenditure for continuous operation and maintenance, market development and other works of the whole system, and can also be used for rule establishment and verification of the AI task newly added in the future, or marking the pre-paid money for data collection and verification.
In the embodiment provided by the invention, the benefits of the proposed expert agreed rule validator and expert agreed rule set-up and updater cannot be bound. If the expert agreed rule verified by the expert agreed rule verifier is wrong, all verifiers of the wrong expert agreed rule are subjected to error punishment. The wrong expert agrees that the Token wallet of the rule verifier will be charged a fee greater than the Token that the contribution of the verification is to take. The specific punishment amount is set according to the actual situation.
In the embodiment provided by the invention, after Token is paid to the AI service system according to the smart contract, the method further comprises the following steps:
s5, writing the transaction information into the blockchain and issuing the transaction information to each accounting node on the blockchain.
In the embodiment provided by the invention, the method further comprises the following steps:
and S6, when the test result is judged to have errors, feeding back to the AI service system.
When the prediction result of the user feedback system is wrong, the prediction result is marked and then stored in the blockchain, and the marked result is forwarded to a marking database. When the number of feedback errors exceeds a certain number of thresholds, the system will check the prediction result data sets to determine whether the corresponding expert agreed rules or data driven model is in error on the user feedback data examples or the user itself is in error. And when the predicted result of the user feedback system is incorrectly fed back, giving a Token reward to the user based on the intelligent contract.
In the embodiment provided by the invention, after Token is paid to the AI service system according to the smart contract, the method further comprises the following steps:
according to the intelligent contracts and the preset expert negotiation rules, the Token is distributed to different types of contributors of the node expert negotiation rules or the data model according to specific proportions or numbers.
Wherein, the expert agreed rule contributors are mainly divided into an establishing or updating person of the expert agreed rule and a verifier of the expert agreed rule.
The data model contributors are mainly divided into algorithm model creators of the data model (updaters are regarded as creators of the updated model), and providers of the annotation data.
In the whole industry AI service system, a conclusion of a certain expert agreed rule or a certain data driving sub-model is adopted by a user, the service corresponds to a Token to be paid, namely, the expert agreed rule or the data driving model is used for paying a corresponding virtual Token, the paid Token is handed to a contributor of the expert agreed rule or the data model of the node according to a specific proportion, and the rest Token is reserved for continuous operation and maintenance, market development and other work of the AI service system, and can also be used for rule establishment and verification of AI tasks newly added in the future or marking pre-paid money of data collection and verification.
The expert agreed rule validator must not bind with the interests of the expert agreed rule contributors. Where the wrong verifier is, his Token wallet will be deducted the amount of Token multiplied by the penalty multiple that the contribution of the verifier corresponds to. The specific penalty factor is set by the platform, in principle greater than 1.
In the embodiment provided by the invention, the roles related to allocating Token mainly comprise:
1) Expert agreements on the rules: a rule proposer or an updater of expert agreed rules including new expert agreed rules, as well as a rule validator that reviews expert agreed rules;
2) Data driven model aspects: including a data provider that provides annotation data, a data validator that reviews annotation data, a model provider or updater that is proposed by an algorithm (model). The annotation data provider can provide annotation feedback for users of the AI service in use, or can simply provide annotation data and obtain rewards without using the AI service.
These roles obtain the corresponding cryptocurrency or are called to obtain the split right of the cryptocurrency according to the contribution degree. Wherein the node expert agrees that the contributors to the rules or data models include rule proponents, model providers, and data providers.
All of the rule or model contribution behaviors (including different types of contributions from different roles), the calling pricing of the rule and model, and all of the user's AI data handling behavior, feedback behavior, and Token transaction behavior are recorded into the accounting node of the blockchain.
In particular, the creation of a data-driven model requires a joint effort of both the initial labeling dataset (data provider) and the algorithmic personnel (model provider). Thus, one data driven model creator is two groups of people. One party is a community of data contributors, occupying a share. The individual virtual red distribution weight of the share is calculated according to the contribution proportion, namely the proportion of the weighted contribution value of each individual contribution and the audit data to the weighted contribution total value of the whole annotation sample data. The other party contributes to the algorithm of the algorithm personnel.
The AI service system is a closed-loop service system, and a user uses the AI service system to automatically form a verified closed loop if the prediction result provided by the system is adopted. The system additionally rewards users for submitting questions, BUGs, and misprediction examples (with the assistance of user verification or subjective belief of the correct output results) of the system, and feeds back to complete the system optimization of the data closed loop. The system will award these feedback to the user a Token award.
The AI service system prohibits the plagiarism specialist from negotiating rules and data driven models. All experts that have proven to have no difference in input and output results agree on rules or data driven models that will automatically mask the following creator according to the first come first go rule. But the same expert agrees on rules, data driven models, and if this is done more quickly, the system automatically agrees on rules based on the partitioning mechanism, giving the updater (i.e., new rule presenter, new model provider, and new data provider) a certain partitioning. The rule creator or model creator enjoys the "patent-like" treatment, prohibiting any plagiarism and imitation for a period of time. Beyond this time period, any newly created expert agrees on rules or models to simulate and operate at a low cost.
Any one of the expert agreed rule auditing work, training the original data provider and verifier, and the subsequent discovery model or expert agreed rule problem providing counter-example gives corresponding Token incentives according to Token price of the corresponding service (the price can be determined by platform internal market bidding or negotiation).
In summary, according to the artificial intelligence service system and the method provided by the invention, the blockchain technology is applied to the professional field, so that the sharing of professional data and expert experience is realized, and the legal rights and interests of industry experts, industry data contributors, algorithm personnel and system maintainers are effectively ensured by a certification mechanism.
The embodiment of the invention also provides a computer readable storage medium. The computer-readable storage medium here stores one or more programs. Wherein the computer readable storage medium may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as read-only memory, flash memory, hard disk, or solid state disk; the memory may also comprise a combination of the above types of memories. The one or more programs, when executed on the computer-readable storage medium, are executable by the one or more processors to perform some or all of the steps of the above-described smart contract and certification-based artificial intelligence service method.
The artificial intelligence service system and the method based on the intelligent contract and the pass provided by the invention are described in detail. Any obvious modifications to the present invention, as would be apparent to those skilled in the art, would constitute an infringement of the patent rights of the invention and would take on corresponding legal liabilities without departing from the true spirit of the invention.
Claims (10)
1. An artificial intelligence service method based on intelligent contracts and passes is characterized by comprising the following steps:
Uploading expert agreed rules or data driven models to an assessment system; the method for judging whether the expert agreed rules are new expert agreed rules or not comprises the following substeps: s101, judging whether an old expert agreed rule which is repeated or partially repeated with the uploaded new expert agreed rule exists in the artificial intelligent service system, and if so, turning to the step S102; otherwise, turning to step S103; wherein, repeated or partial repeated refers to that the applicable limiting conditions of the input data are the same or partially overlapped; s102, comparing with the old expert agreed rules, judging whether the new expert agreed rules are updated, if so, turning to a step S103; otherwise, refusing the uploading of the new expert agreed rules; s103, expert verification is carried out on the uploaded new expert negotiation rules through the following substeps: determining expert committee members for auditing according to the field to which the uploaded expert negotiation rules belong; setting a verification threshold, and when the number of expert committee members qualified by the verification uploading expert negotiation rule is larger than the verification threshold, the uploading expert negotiation rule is qualified and the new expert negotiation rule is released; otherwise, rejecting the expert agreed rule to issue to the artificial intelligence service system; or, performing conflict test on the uploaded data driving model to judge whether the model is a new data driving model, including the following sub-steps: s111, carrying out predictive test on the labeling data set continuously accumulated by the system by the original data driving model and the newly added data driving model at the same time, and labeling the consistency of the new data driving model and each old model one by one; s112, judging whether an existing old model exists or not by the evaluation system, wherein the output of the old model is completely consistent with the output of the new model; if yes, rejecting the release of the new data driving model; if no such model exists, the process goes to step S113; s113, the evaluation system evaluates and verifies the uploaded new data driving model through the following substeps: the evaluation system calculates the false report rate and the false report rate of the uploaded data driving model based on the corresponding marked sample data set; multiplying the false alarm rate and the false alarm rate by corresponding cost functions to obtain a final loss expected value, and if the final loss expected value of the new data driving model is smaller than the final loss expected value of the original data driving model and is lower than a model prediction loss function threshold preset by a system, verifying the new data driving model;
The expert agreed rules or data driven models passing verification are stored in a mode library and are issued to the corresponding artificial intelligent service system; simultaneously binding a benefit distribution mechanism to an intelligent contract, and writing the benefit distribution mechanism into a blockchain;
when the input data call the artificial intelligent service, expert negotiation rules or data driving models applicable to the artificial intelligent service system call mode library obtain a prediction result;
after the user selects the prediction result, paying the pass to the artificial intelligence service system according to the intelligent contract;
wherein all experts proved to have no difference in input and output results agree on rules or data driven models to mask the following creator according to first-come first-served rules; the same expert agrees that rules or data driven models, if implemented faster, give the updater a predetermined split based on the split mechanism and prohibit any plagiarism and imitation for a certain period of time; beyond this period of time, the expert agrees on rules or that the data driven model may simulate and operate at a low cost.
2. The artificial intelligence service method according to claim 1, wherein:
it is determined whether the expert negotiation rules update one or more of whether the input including the new expert negotiation rules adds a new data object, whether the new expert negotiation rules further subdivide the input old data object, whether the input object of the new expert negotiation rules adds an attribute, whether the new expert negotiation rules modify an attribute, whether the new expert negotiation rules further divide the old expert negotiation rules, and whether a condition or parameter in the rules changes.
3. The artificial intelligence service method according to claim 1, wherein:
when the expert committee member examines the expert agreed rules uploaded by the expert, the branch sub-rules are established by correcting the uploaded expert agreed rules, and the expert establishing the branch sub-rules is the new rule owner.
4. The artificial intelligence service method of claim 1, wherein when the input data invokes the artificial intelligence service for prediction, expert negotiation rules or data driven models applicable in the artificial intelligence service system invocation mode library obtain a prediction result and output, comprising the steps of:
when input data call artificial intelligent service to predict, the artificial intelligent service system searches corresponding expert negotiation rules or data driving models according to the input data;
judging the input data according to the found expert agreed rules or the data driving model to obtain a prediction result;
and outputting and displaying the selected prediction result.
5. The artificial intelligence service method according to claim 1, wherein:
when the selected prediction result corresponds to a plurality of expert agreed rules or data driven models, the artificial intelligence service system matches the expert agreed rules or data driven models which select the lowest transaction price.
6. The artificial intelligence service method according to claim 1, wherein:
when the number of times of the predicted result errors of the user feedback system exceeds a specific number threshold, checking the predicted result data sets to judge whether the corresponding expert agreed rules or the data driving model generate errors on the data examples fed back by the users or whether the users judge the errors by themselves.
7. The artificial intelligence service method according to claim 6, further comprising the steps of:
and when the predicted result of the user feedback system is incorrectly fed back, the user is given a certification reward based on the intelligent contract.
8. The artificial intelligence service method according to claim 1, further comprising the steps of:
pass is assigned to different types of contributors according to smart contracts and pre-established expert agreed rules and recorded into the accounting nodes of the blockchain.
9. An artificial intelligence service system based on intelligent contracts and passes for realizing the artificial intelligence service method of any one of claims 1-8, which is characterized by comprising a knowledge uploading module, a knowledge verifying module, a publishing module, a predicting module and a payment module;
The knowledge uploading module is used for uploading expert agreed rules or a data driving model to the evaluation system;
the knowledge verification module is used for verifying the knowledge transmitted by the knowledge uploading module and transmitting expert agreed rules or data driven models passing verification to the issuing module;
the issuing module is used for binding the authenticated expert agreed rules or the evaluation systems corresponding to the data driving models to intelligent contracts, writing block chains and issuing to the artificial intelligent service system;
when user input data call artificial intelligent service to predict, the prediction module predicts according to expert agreed rules or data driving models issued by the issuing module and outputs a prediction result to the user;
the payment module pays the pass to the artificial intelligence service system from the user account according to the intelligent contract according to the prediction result of the prediction module selected by the user, and issues transaction information through the issuing module;
wherein all experts proved to have no difference in input and output results agree on rules or data driven models to mask the following creator according to first-come first-served rules; the same expert agrees that rules or data driven models, if implemented faster, give the updater a predetermined split based on the split mechanism and prohibit any plagiarism and imitation for a certain period of time; beyond this period of time, the expert agrees that rules or data driven models are modeled and run at a low cost.
10. The artificial intelligence service system of claim 9, further comprising a collision detection module; the conflict detection module is used for carrying out conflict detection on the uploaded expert agreed rules or the data driving model and judging whether the uploaded expert agreed rules or the uploaded data driving model are new expert agreed rules or new data driving models.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810901731.1A CN109191293B (en) | 2018-08-09 | 2018-08-09 | Artificial intelligence service system and method based on intelligent contract and general evidence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810901731.1A CN109191293B (en) | 2018-08-09 | 2018-08-09 | Artificial intelligence service system and method based on intelligent contract and general evidence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109191293A CN109191293A (en) | 2019-01-11 |
| CN109191293B true CN109191293B (en) | 2023-11-14 |
Family
ID=64921183
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810901731.1A Active CN109191293B (en) | 2018-08-09 | 2018-08-09 | Artificial intelligence service system and method based on intelligent contract and general evidence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109191293B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110222721B (en) * | 2019-05-10 | 2021-07-30 | 达闼机器人有限公司 | Data processing method, data processing device, block chain node and storage medium |
| CN110276613B (en) * | 2019-06-20 | 2020-03-27 | 卓尔智联(武汉)研究院有限公司 | Block chain-based data processing apparatus, method, and computer-readable storage medium |
| CN110533419B (en) * | 2019-07-24 | 2023-09-26 | 创新先进技术有限公司 | Block chain-based sharing method, device and server of wind control model |
| CN111062705A (en) * | 2019-12-25 | 2020-04-24 | 东华理工大学 | Online English education big data platform |
| CN111526180B (en) * | 2020-03-30 | 2022-05-24 | 中国建设银行股份有限公司 | Data sharing method and device |
| CN111901753A (en) * | 2020-06-04 | 2020-11-06 | 中国联合网络通信集团有限公司 | Infectious disease close contact person determination method based on block chain and block chain link point |
| CN117235782A (en) * | 2023-08-31 | 2023-12-15 | 北京可利邦信息技术股份有限公司 | Method, system and terminal for realizing privacy calculation data security based on alliance chain |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101093559A (en) * | 2007-06-12 | 2007-12-26 | 北京科技大学 | Method for constructing expert system based on knowledge discovery |
| CN107194587A (en) * | 2017-05-24 | 2017-09-22 | 暨南大学 | It is a kind of to be circulated based on block chain and the art work of expert system and identification register method |
| CN107240018A (en) * | 2017-07-25 | 2017-10-10 | 成都励睿德企业管理有限公司 | A kind of method and system for being used to pay tranaction costs in block chain network |
| CN107563905A (en) * | 2017-07-20 | 2018-01-09 | 西安电子科技大学 | A kind of academic platform service system and method for building up based on block chain |
-
2018
- 2018-08-09 CN CN201810901731.1A patent/CN109191293B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101093559A (en) * | 2007-06-12 | 2007-12-26 | 北京科技大学 | Method for constructing expert system based on knowledge discovery |
| CN107194587A (en) * | 2017-05-24 | 2017-09-22 | 暨南大学 | It is a kind of to be circulated based on block chain and the art work of expert system and identification register method |
| CN107563905A (en) * | 2017-07-20 | 2018-01-09 | 西安电子科技大学 | A kind of academic platform service system and method for building up based on block chain |
| CN107240018A (en) * | 2017-07-25 | 2017-10-10 | 成都励睿德企业管理有限公司 | A kind of method and system for being used to pay tranaction costs in block chain network |
Non-Patent Citations (2)
| Title |
|---|
| 卢力行."基于知识管理的专家资源管理平台的设计与实现".《中国优秀硕士学位论文全文数据库（信息科技辑）》.2015,正文第21-32页. * |
| 薛腾飞 等."基于区块链的医疗数据共享模型研究"."基于区块链的医疗数据共享模型研究".2017,第第43卷卷(第第43卷期),第1555-1562页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109191293A (en) | 2019-01-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109191293B (en) | Artificial intelligence service system and method based on intelligent contract and general evidence | |
| US20210357029A1 (en) | Ai solution selection for an automated robotic process | |
| US11552785B2 (en) | Methods and systems for a synchronized distributed data structure for federated machine learning | |
| Fischer et al. | The good, the bad, and the unfamiliar: The challenges of reputation formation facing new firms | |
| JP2023513146A (en) | Selection and configuration of artificial intelligence | |
| US11393038B2 (en) | Machine structured plan description | |
| Ghili | Network formation and bargaining in vertical markets: The case of narrow networks in health insurance | |
| CN112735549A (en) | Data processing method and data processing system based on medical insurance data | |
| US11393039B2 (en) | Machine structured plan description | |
| Zhang et al. | Feedback reviews and bidding in online auctions: An integrated hedonic regression and fuzzy logic expert system approach | |
| Pfeiffer et al. | Algorithmic Fairness in AI: An Interdisciplinary View | |
| Scantamburlo | Non-empirical problems in fair machine learning | |
| Brown et al. | Information and Disparities in Health Care Quality: Evidence from GP Choice in England | |
| Bao et al. | A methodology for supplier selection under the curse of dimensionality problem based on fuzzy quality function deployment and interval data envelopment analysis | |
| Erkal et al. | Ethical evaluation of artificial intelligence applications in radiotherapy using the Four Topics Approach | |
| Kamal et al. | Debiased machine learning for estimating the causal effect of urban traffic on pedestrian crossing behavior | |
| Woldesenbet | Highway infrastructure data and information integration & assessment framework: A data-driven decision-making approach | |
| Burdett et al. | The efficacy of utility functions for multicriteria hospital case‐mix planning | |
| Jing et al. | Mining personalized individual semantics of self-confidence participants in linguistic group decision-making | |
| Sreeram et al. | Short-term Forecasting Ability of Hybrid Models for BRIC Currencies | |
| Lakkaraju | Human-Centric Machine Learning: Enabling Machine Learning for High-Stakes Decision-Making | |
| US20240006061A1 (en) | Data currency system for digital human healthcare and medical data exchange, analytics, and applications | |
| Püschel et al. | Comparison of Discrete Choice and Machine Learning Models for Simultaneous Modeling of Mobility Tool Ownership in Agent-Based Travel Demand Models | |
| Asemi et al. | Adaptive neuro-fuzzy inference system for customizing investment type based on the potential investors’ demographics and feedback | |
| Yang | Making Decisions Under High Stakes: Trustworthy and Expressive Bayesian Deep Learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: Room 1101-61, No. 379 Tayuan Road, Suzhou High-tech Zone, Suzhou City, Jiangsu Province 215011 Applicant after: Suzhou Navitas Space-Time Information Technology Co.,Ltd. Address before: Room 1-403, Block B, Qilu Software Park Building, No. 1768 Xinluo Street, High tech Zone, Jinan City, Shandong Province, 250101 Applicant before: JINAN NAWEI INFORMATION TECHNOLOGY Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |