CN109902506A - A kind of local difference private data sharing method and system of more privacy budgets - Google Patents
A kind of local difference private data sharing method and system of more privacy budgets Download PDFInfo
- Publication number
- CN109902506A CN109902506A CN201910015907.8A CN201910015907A CN109902506A CN 109902506 A CN109902506 A CN 109902506A CN 201910015907 A CN201910015907 A CN 201910015907A CN 109902506 A CN109902506 A CN 109902506A
- Authority
- CN
- China
- Prior art keywords
- privacy
- client
- data
- server
- factor
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of local difference private data sharing method of more privacy budgets and systems, propose under local difference privacy application scenarios, the frame of the privacy budget factor is determined by user oneself.The present invention is on existing local difference privacy protocol basis, it is added to the step of allowing user to first confirm that itself privacy factor, data collector obtains a point bucket according to the distribution for the privacy factor that all users submit, user's foundation divides barrel knot fruit shape at group, the noisy data that the user of each group finally submits will use the identical privacy factor, the privacy factor can be less than or equal to the Original submission privacy factor of user, to meet the privacy requirements of user.Data collector calculates more accurately statistical result by the methods of maximum likelihood, and its accuracy can obtain mathematical proof.The present invention can be applied directly to existing local difference privacy algorithm such as RAPPOR, in SH agreement, it is easy to accomplish.
Description
Technical field
The invention belongs to computer science and field of information security technology, in local difference privacy application scenarios, propose
The secret protection rank that family independently defines itself can be used in a kind of frame independently confirming privacy budget factor ε by user,
It can effectively prevent data collector's malice of some untrusteds from extracting the true private data of user using high budget factor ε.
It can be proved that the present invention can be used in the statistical activity that the local difference privacy agreement of standard is carried out, and accuracy rate is high, real
It is strong with property.
Background technique
By stringent mathematical proof, difference privacy is one of current strongest privacy guarantees law.Its principle is to use
Carefully adjusted noise covers the data of user.When many users submit respective sensitive data, data center is according to right
The statistical result of all data adds noise into data and reaches balance, and generates significant information.But this traditional difference
The original sensitive data of user is focused on a data center by point privacy technology, then is added by data center to user data
Work, thus have here one it is very crucial it is assumed that i.e. data center/data collector be it is believable, the hidden of user will not be revealed
It is private.Obvious this premise exists only in theory, though the subjective wish without revealing privacy of user of data collector, but by
In external factors such as network attacks, the private data of user is equally possible to be obtained illegally.
And local difference privacy directly can carry out privacy to sensitive data in user client (mobile phone app, browser)
Change processing, data center is also unable to get the original sensitive data of user, to prevent data center or data collector lets out
Reveal the possibility of privacy of user.Local difference privacy has an important privacy budget factor ε, it is represent protection to a certain degree
Security level possessed by data afterwards.When ε is bigger, the privacy of data is lower, while its availability also can be higher.Working as
In modern local difference privacy application (Rappor agreement, SH agreement), the setting of ε numerical value is determined by data collector substantially, this
The safety that represent user's data is still not exclusively determined by itself, and local difference privacy is caused to be deposited in application process
The loophole of excessive ε is chosen in data collector's malice.Therefore, determine that the safety of oneself sensitive data is one by user itself
Urgent demand.
Summary of the invention
It is an object of the invention in local difference privacy application scenarios, realize that one kind determines that privacy is pacified by user itself
Full grade method for distinguishing.This method is based on maximum likelihood method and sampling theory, can be integrated into existing local difference privacy association
View is as in Rappor, SH.The present invention can also be embodied directly at server-side and client two, and addition is supported pre- from selecting respectively
Calculate the module of the factor.
The technical solution adopted by the invention is as follows:
A kind of local difference private data sharing method of more privacy budgets, comprising the following steps:
Client receives Data Collection task from server-side；
Client call local difference privacy algorithm, the privacy budget factor pair sensitive data defined using itself are disturbed
It is dynamic；
Secure data after disturbance is sent to server-side by trusted channel by client.
Further, the client determines the total privacy budget factor of itself according to the privacy demand of itself, every time visitor
Family end submits the data of local difference secret protection that can consume a part of total privacy budget factor, to control sensitive data submission
Number.
Further, the client and the server-side negotiate to determine the privacy budget of actual use by following steps
The factor:
The sensitive data that client possesses itself assigns the corresponding consumable maximum privacy budget factor of single institute；
Client finds the corresponding maximum privacy budget factor ε max of sensitive data, ε max is submitted according to task type
To server-side, so that server-side confirms the distribution for the privacy budget factor that all users submit, it is according to Sampling that its is hidden
Private budget factor value range divides bucket, then the lower limit ε ' of bucket is divided to return to client corresponding to the ε max by user；
Client receives ε ', as the privacy budget factor of client actual use from server-side.
A kind of local difference private data sharing method of more privacy budgets, includes the following steps
Server-side issues Data Collection task to client；
Server-side receives client by calling local difference privacy algorithm, the privacy budget factor pair defined using itself
Sensitive data disturbed after secure data；
Server-side according to the statistical result for from the secure data after the received disturbance of client, obtaining Data Collection task and
Accuracy.
Further, the server-side and the client negotiate to determine that the client is actually used by following steps
The privacy budget factor:
Server-side receives the corresponding maximum privacy budget factor ε max of sensitive data from client；
Server-side confirms the distribution for the privacy budget factor that all clients are submitted, according to Sampling by its privacy budget
Factor value range divides bucket, and the lower limit ε ' of bucket is then divided to return to client corresponding to the ε max by user, as client
The privacy budget factor of actual use.
Further, client is grouped by server-side according to the privacy budget factor that it is used, and is calculated separately to each group
Then statistical result is merged all results using maximum likelihood method.
A kind of client comprising:
Privacy budget factor computing module is responsible for determining the total privacy budget factor of itself according to the privacy demand of itself,
Each client submits the data of local difference secret protection that can consume a part of total privacy budget factor, to control sensitive number
According to submission number；
Local difference privacy algoritic module is responsible for storing local difference privacy algorithm, by the sensitive data use pair of user
The budget factor answered is disturbed, and the secure data after disturbance is then sent to server-side by trusted channel.
A kind of server-side comprising:
Privacy budget factor statistics and negotiation module are responsible for negotiating to determine that the client is actually used with the client
The privacy budget factor, and be sent to the client；
It is grouped disturbance information statistical module, is responsible for after the secure data after being disturbed from the client, by most
The method of maximum-likelihood obtains the final statistical result of privacy of user data.
Further, the server-side further includes privacy algorithms selection module, is responsible for that client is made to pass through actual use
The privacy budget factor confirms the local difference privacy algorithm optimized accordingly.
A kind of local difference private data share system of more privacy budgets, including client recited above and service
End.
Compared to the prior art, the advantages of the present invention are mainly reflected in:
1) make the budget factor in the customized local difference privacy of user, realize respectively different privacy demands, can make more
More users participates in the data collection activity of local difference privacy.
2) distribution for the budget factor that the order of accuarcy of statistical result can be provided according to all users accurately calculates out.
3) data collector only adds on the basis of original local difference privacy algorithm (such as RAPPOR, SH agreement)
A small amount of operation such as maximum likelihood algorithm, can support the privacy demand that user is different, easy to accomplish.
Detailed description of the invention
Fig. 1 is the local difference privacy algorithm configuration diagram of the customized privacy budget factor of user.
Specific embodiment
In order to make the foregoing objectives, features and advantages of the present invention clearer and more comprehensible, below by specific embodiment and
Attached drawing is described in further details the present invention.
The present invention is made of client and server-side two parts:
One, client
Client is located in the equipment of user, can be mobile phone app, browser plug-in etc..The original private data of user
It is just stored only in client, such user can grasp the privacy-sensitive data of itself completely.When needing to share these data
When, user can also call local difference privacy algorithm, and the privacy budget factor ε defined using itself disturbs the data
It is dynamic, the data of the safety after disturbance are finally sent to data collector by trusted channel.
The structure composed of the technical solution of client mainly includes privacy budget factor computing module, original referring to attached drawing 1
Sensitive data collection modules, local difference privacy algoritic module.Privacy budget factor computing module is responsible for control sensitive data and is mentioned
Number is handed over, each user submits the data of local difference secret protection, can all consume a part of total privacy budget factor.It is original quick
Sense data aggregation module stores the initial data and the corresponding security level of Various types of data of user.Local difference privacy is calculated
Method module can store a variety of local difference privacy algorithms, be responsible for disturbing the data of user using the randomization of the corresponding budget factor
It is dynamic, the data after disturbance are then sent to data collector.
Two, server-side
Server-side is controlled by data collector, and the purpose of data collector is the statistics letter for obtaining all privacy of user data
Breath (the discrete distribution of such as frequent episode, marginal probability statistics etc.), the local difference privacy that statistical information can be provided by user
Noisy data after algorithm protection is calculated, but the order of accuarcy of this statistical information depends on the local difference that user uses
The size of the budget factor of privacy algorithm.In the present invention, the budget factor of user is determined by user itself.
The structure composed of the technical solution of server-side is referring to attached drawing 1, privacy budget factor statistics and negotiation module mistake herein
It is responsible for the distribution map that statistics total user submits used ε numerical value before noisy data in journey, distribution map is usually continuously distributed
, so the module continues to divide bucket to become discrete values distribution map, dividing the foundation of bucket is Sampling, makes the use in each bucket
Amount mesh, which can more to reach, individually implements primary local difference privacy statistics.Divide after bucket, by the lower limit of the bucket where user
ε ' is returned to user, and ε ' is just as negotiation result (the privacy budget factor that client finally uses).It is grouped disturbance information statistics
Module obtains final statistical result by the method for maximum likelihood after the noisy data for obtaining user's submission.As shown in Figure 1,
Privacy algorithms selection module can also be further arranged in server-side, allow server-side according to the value size requirements user of ε '
Specific local difference privacy agreement is selected to do disturbance of data, it means that ε ' different users, the Perturbed algorithms used
Difference, to improve the accuracy rate of final statistical result.
Example explanation will be done to the specific implementation of key technology module described in summary of the invention below, but not with this
Kind explains the range of limitation invention.
This basic module of local difference privacy algoritic module is introduced first.Local difference privacy algoritic module position
In user client, the budget factor that the sensitive data and user for inputting user define it, the number protected after output disturbance
According to.The present invention needs the partial function provided using it, however the implementation of this module itself does not then consider model in the present invention
In enclosing, existing local difference privacy agreement substitution can be used.The present invention uses rappor protocol parameter h=1 in this example
This local difference privacy agreement do more intuitive explanation, statistical result is the distribution situation of the private data of user.
Below by explain the functions of the present invention.
1. the main flow of pair technology is illustrated:
1.1) user confirms the privacy budget factor
User determines the total privacy budget factor of itself according to the privacy demand of itself.And the sensitivity that itself is possessed
Data assign the corresponding consumable maximum privacy budget factor of single institute.
1.2) it collects tasks secure rank and confirms process
1) data collector issues a collection task in server-side, and task corresponds to the sensitive number that required user submits
According to.
2) user finds the corresponding maximum privacy budget factor ε max of the sensitive data in client according to task type,
ε max is first submitted into gatherer.
3) server-side confirms the distribution for the privacy budget factor that all users submit, according to Sampling by its privacy budget
Factor value range divides bucket, and the Sampling used herein mainly determines a suitable sample size, makes in each point of bucket
The number for the user for including be big enough to the user in this point of bucket individually carry out local difference privacy statistics the result is that effective
's.Then the lower limit ε ' of bucket is divided to return to user corresponding to the ε max by user, (ε ' < ε max).In the present invention, user's foundation
Divide barrel knot fruit shape at group, the noisy data that the user of each group finally submits will use the identical privacy budget factor, should
The privacy budget factor can be less than or equal to the Original submission privacy budget factor of user, to meet the privacy requirements of user.
4) further optimization method, server-side is according to the size of ε ' and the relationship (frequent episode of statistics task special parameter
The candidate spatial size d) set when statistics, the user for keeping ε ' different use different Perturbed algorithms as local difference privacy agreement
Main body.Such as in frequent episode statistics, as d > 3eε′When+2, select OptimalSchemes agreement that statistical result can be made accurate
Degree can be optimal, otherwise d < 3eε′KRR agreement is selected when+2.This step is an optimization method, specifically need to be according to actual
Statistics task determines, is still analyzed in subsequent instruction using rappor agreement.
5) data collector confirms that current collection task institute is attainable total accurate according to the result of the lower limit divided after bucket
Degree.Accuracy refers to the variance of statistical result；It is obtained using the method for maximum likelihood by the lower limit ε joint account of all points of buckets
Population variance, as overall accuracy.
1.3) local difference privacy statistical flowsheet
1) data after user receives new ε ', after being disturbed using local difference privacy algorithm.If used
Rappor agreement, the private data of user are x1, x1 can be converted to again to one-hot coding first, then to the every of coding
One with probability 1/eε′/2Probability negate, obtain the data of disturbance.Client sends it to data collector, and at itself
Total privacy budget factor computing module in, subtract the budget of ε ', the work of client ends here.
2) server-side will be flocked together after negotiation using the user of the identical privacy budget factor, that is, is grouped, to every
A group respectively counting statistics as a result, then all results are merged using maximum likelihood method.
The method of " to each group of difference counting statistics result " is:
When server-side gets noisy data p (x, the ε ') of a group user (budget that the user of the group uses because
Son is all ε '), and p (x, ε ') it is binary vector format (such as { 0,1,0,1 ..., 0 }), the noisy data of all users is tired out first
It adds up, forms an array (such as { 10210,5214,26842 ..., 22358 }).Then using the warp of local difference privacy
Allusion quotation homing method obtains estimated value (such as { 7%, 3%, 15% ..., 14% }), and estimated value is exactly point of the private data of user
Cloth situation, it is a unbiased esti-mator, has corresponding variance.The user of each group is corresponding to generate above-mentioned estimation
Value, they are all a Sampling Estimations of final statistical result.
2. the more privacy budget factor disturbance results of server-side merge algorithm
Since local difference privacy algorithm is established on the basis of stochastic phase (random response) algorithm,
So the estimated result that local difference privacy algorithm counts is a unbiased esti-mator, and this unbiased esti-mator can be calculated
Variance size.So the input and output for merging algorithm are as follows:
Input: server-side by user after the grouping of the privacy budget factor (assuming that M group), the statistical result that has (f1,
F2 ... fM)；The privacy budget factor and number of every group of user.((ε1, n1), (ε2, n2) ..., (εm, nM))。
Output: the statistical result finally merged
The characteristic of unbiased esti-mator is utilized in the algorithm, uses the different group of the method combination budget factor of maximum likelihood
User's statistical result calculates variance size with the method for mathematics and shows its accuracy rate with higher.
Specific merging process is as follows:
1) the accuracy i.e. variance size of the estimation of each group is calculated separately, the variance of different agreement is of different sizes.Make
When with rappor agreement,
2) coefficient used when merging is calculated,
3) merging formula is
4) it is calculated using the method for maximum likelihoodPopulation variance:
The above embodiments are merely illustrative of the technical solutions of the present invention rather than is limited, the ordinary skill of this field
Personnel can be with modification or equivalent replacement of the technical solution of the present invention are made, without departing from the principle and scope of the present invention, originally
The protection scope of invention should be subject to described in claims.
Claims (10)
1. a kind of local difference private data sharing method of more privacy budgets, which comprises the following steps:
Client receives Data Collection task from server-side；
Client call local difference privacy algorithm, the privacy budget factor pair sensitive data defined using itself are disturbed；
Secure data after disturbance is sent to server-side by trusted channel by client.
2. the method according to claim 1, wherein the client determines itself according to the privacy demand of itself
Total privacy budget factor, each client submit the data of local difference secret protection can consume a part of total privacy budget because
Son, so that controlling sensitive data submits number.
3. method according to claim 1 or 2, which is characterized in that client and server-side are negotiated really by following steps
Surely the privacy budget factor actually used:
The sensitive data that client possesses itself assigns the corresponding consumable maximum privacy budget factor of single institute；
Client finds the corresponding maximum privacy budget factor ε max of sensitive data, ε max is submitted to clothes according to task type
Business end, so that server-side confirms the distribution for the privacy budget factor that all users submit, it is according to Sampling that its privacy is pre-
It calculates factor value range and divides bucket, then the lower limit ε ' of bucket is divided to return to client corresponding to the ε max by user；
Client receives ε ', as the privacy budget factor of client actual use from server-side.
4. a kind of local difference private data sharing method of more privacy budgets, which is characterized in that include the following steps
Server-side issues Data Collection task to client；
Server-side receives client by calling local difference privacy algorithm, and the privacy budget factor pair using itself definition is sensitive
Data disturbed after secure data；
Server-side is according to the statistical result for from the secure data after the received disturbance of client, obtaining Data Collection task and accurate
Degree.
5. according to the method described in claim 4, it is characterized in that, server-side and client negotiate to determine institute by following steps
State the privacy budget factor of client actual use:
Server-side receives the corresponding maximum privacy budget factor ε max of sensitive data from client；
Server-side confirms the distribution for the privacy budget factor that all clients are submitted, according to Sampling by its privacy budget factor
Value range divides bucket, and the lower limit ε ' of bucket is then divided to return to client corresponding to the ε max by user, as client reality
The privacy budget factor used.
6. according to the method described in claim 4, it is characterized in that, the server-side privacy budget that uses client according to it because
Subgroup, to each group of difference counting statistics as a result, then being merged all results using maximum likelihood method.
7. a kind of client characterized by comprising
Privacy budget factor computing module is responsible for determining the total privacy budget factor of itself according to the privacy demand of itself, every time
Client submits the data of local difference secret protection that can consume a part of total privacy budget factor, mentions to control sensitive data
Hand over number；
Local difference privacy algoritic module is responsible for storing local difference privacy algorithm, the sensitive data of user be used corresponding
The budget factor is disturbed, and the secure data after disturbance is then sent to server-side by trusted channel.
8. a kind of server-side characterized by comprising
Privacy budget factor statistics and negotiation module are responsible for negotiating to determine that the client is real with client described in claim 8
The privacy budget factor that border uses, and it is sent to the client；
It is grouped disturbance information statistical module, is responsible for after the secure data after being disturbed from the client, seemingly by maximum
Right method obtains the final statistical result of privacy of user data.
9. server-side according to claim 8, which is characterized in that further include privacy algorithms selection module, be responsible for making client
End confirms the local difference privacy algorithm optimized accordingly by the privacy budget factor of actual use.
10. a kind of local difference private data share system of more privacy budgets, which is characterized in that including described in claim 7
Client and claim 8 or 9 described in server-side.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910015907.8A CN109902506B (en) | 2019-01-08 | 2019-01-08 | Local differential privacy data sharing method and system with multiple privacy budgets |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910015907.8A CN109902506B (en) | 2019-01-08 | 2019-01-08 | Local differential privacy data sharing method and system with multiple privacy budgets |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109902506A true CN109902506A (en) | 2019-06-18 |
| CN109902506B CN109902506B (en) | 2021-02-26 |
Family
ID=66943725
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910015907.8A Active CN109902506B (en) | 2019-01-08 | 2019-01-08 | Local differential privacy data sharing method and system with multiple privacy budgets |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109902506B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111090877A (en) * | 2019-11-21 | 2020-05-01 | 腾讯科技（深圳）有限公司 | Data generation method, data acquisition method, corresponding devices and storage medium |
| CN111241582A (en) * | 2020-01-10 | 2020-06-05 | 鹏城实验室 | Data privacy protection method and device and computer readable storage medium |
| CN111669366A (en) * | 2020-04-30 | 2020-09-15 | 南京大学 | Localized differential private data exchange method and storage medium |
| CN113297621A (en) * | 2021-06-10 | 2021-08-24 | 南京航空航天大学 | Data collection method based on personalized local differential privacy |
| CN114154202A (en) * | 2022-02-09 | 2022-03-08 | 支付宝(杭州)信息技术有限公司 | Wind control data exploration method and system based on differential privacy |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095447A (en) * | 2015-07-24 | 2015-11-25 | 武汉大学 | Distributed w-event differential privacy infinite streaming data distribution method |
| CN106991335A (en) * | 2017-02-20 | 2017-07-28 | 南京邮电大学 | A kind of data publication method based on difference secret protection |
| CN107302521A (en) * | 2017-05-23 | 2017-10-27 | 全球能源互联网研究院 | The sending method and method of reseptance of a kind of privacy of user data |
| CN107689950A (en) * | 2017-06-23 | 2018-02-13 | 平安科技（深圳）有限公司 | Data publication method, apparatus, server and storage medium |
| CN108234493A (en) * | 2018-01-03 | 2018-06-29 | 武汉大学 | The space-time crowdsourcing statistical data dissemination method of secret protection under insincere server |
-
2019
- 2019-01-08 CN CN201910015907.8A patent/CN109902506B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095447A (en) * | 2015-07-24 | 2015-11-25 | 武汉大学 | Distributed w-event differential privacy infinite streaming data distribution method |
| CN106991335A (en) * | 2017-02-20 | 2017-07-28 | 南京邮电大学 | A kind of data publication method based on difference secret protection |
| CN107302521A (en) * | 2017-05-23 | 2017-10-27 | 全球能源互联网研究院 | The sending method and method of reseptance of a kind of privacy of user data |
| CN107689950A (en) * | 2017-06-23 | 2018-02-13 | 平安科技（深圳）有限公司 | Data publication method, apparatus, server and storage medium |
| CN108234493A (en) * | 2018-01-03 | 2018-06-29 | 武汉大学 | The space-time crowdsourcing statistical data dissemination method of secret protection under insincere server |
Non-Patent Citations (1)
| Title |
|---|
| 刘慧艺: "《面向数据流的差分隐私直方图发布》", 《中国优秀硕士学位论文全文数据库-信息科技辑》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111090877A (en) * | 2019-11-21 | 2020-05-01 | 腾讯科技（深圳）有限公司 | Data generation method, data acquisition method, corresponding devices and storage medium |
| CN111090877B (en) * | 2019-11-21 | 2023-07-28 | 腾讯科技（深圳）有限公司 | Data generation and acquisition methods, corresponding devices and storage medium |
| CN111241582A (en) * | 2020-01-10 | 2020-06-05 | 鹏城实验室 | Data privacy protection method and device and computer readable storage medium |
| CN111241582B (en) * | 2020-01-10 | 2022-06-10 | 鹏城实验室 | Data privacy protection method and device and computer readable storage medium |
| CN111669366A (en) * | 2020-04-30 | 2020-09-15 | 南京大学 | Localized differential private data exchange method and storage medium |
| CN111669366B (en) * | 2020-04-30 | 2021-04-27 | 南京大学 | Localized differential private data exchange method and storage medium |
| CN113297621A (en) * | 2021-06-10 | 2021-08-24 | 南京航空航天大学 | Data collection method based on personalized local differential privacy |
| CN114154202A (en) * | 2022-02-09 | 2022-03-08 | 支付宝(杭州)信息技术有限公司 | Wind control data exploration method and system based on differential privacy |
| CN114154202B (en) * | 2022-02-09 | 2022-06-24 | 支付宝(杭州)信息技术有限公司 | Wind control data exploration method and system based on differential privacy |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109902506B (en) | 2021-02-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109902506A (en) | A kind of local difference private data sharing method and system of more privacy budgets | |
| US11716266B2 (en) | Network security monitoring and correlation system and method of using same | |
| JP6261665B2 (en) | Determining connections within a community | |
| CN108351881A (en) | The system and method for optimally in distributed mode | |
| US9320006B2 (en) | Categorized location identification based on historical locations of a user device | |
| CN109600798A (en) | Multi-domain resource allocation method and device in a kind of network slice | |
| US11386663B1 (en) | Reference-free system for determining quality of video data | |
| US20180213047A1 (en) | System and method to determine the presence status of a registered user on a network | |
| CN116685985A (en) | Joint learning system and method with diversified feedback | |
| CN112352412B (en) | Network traffic processing method and device, storage medium and computer equipment | |
| CN114884682B (en) | Crowd sensing data stream privacy protection method based on self-adaptive local differential privacy | |
| WO2020087879A1 (en) | Method, device, and system for protecting private information | |
| CN107302521A (en) | The sending method and method of reseptance of a kind of privacy of user data | |
| CN107659505B (en) | SDN network routing method and SDN controller | |
| Singh et al. | Service Versus Protection: a Bayesian learning approach for trust provisioning in edge of things environment | |
| CN110874638B (en) | Behavior analysis-oriented meta-knowledge federation method, device, electronic equipment and system | |
| CN108171570A (en) | A kind of data screening method, apparatus and terminal | |
| WO2019047345A1 (en) | Method of generating one-time password sending policy, and one-time password sending method | |
| EP2622499B1 (en) | Techniques to support large numbers of subscribers to a real-time event | |
| CN101841529A (en) | Privacy information protection method based on informationism and trust | |
| KR101609056B1 (en) | Smart context distribution framework based on messaging service for internet of things and control method for the same | |
| CN114723071B (en) | Federal learning method and device based on client classification and information entropy | |
| CN106570699A (en) | Client contact information excavation method and server | |
| Huang | The value-of-information in matching with queues | |
| WO2023015165A1 (en) | Configuration of user interface for risk assessment visualization |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |