Download to read offline
![IRTModelsFitPrivacy-AwareCloudSharing?
• Ex: With which privacy level would you
share a project presentation with a friend?
• Standardized Infit Statistic:
• (x-axis values should lie in [-2,2])
16
81
96
Dichotomous case
Sensitivity
Infit t-statistic
A dot
represents a
context](https://image.slidesharecdn.com/onlineversion-160710210747/85/C3P-Context-Aware-Crowdsourced-Cloud-Privacy-at-PETS-2014-52-320.jpg)
![IRTModelsFitPrivacy-AwareCloudSharing?
• Ex: With which privacy level would you
share a project presentation with a friend?
• Standardized Infit Statistic:
• (x-axis values should lie in [-2,2])
16
81
96
Dichotomous case
Sensitivity
Infit t-statistic
A dot
represents a
context](https://image.slidesharecdn.com/onlineversion-160710210747/85/C3P-Context-Aware-Crowdsourced-Cloud-Privacy-at-PETS-2014-53-320.jpg)
![IRTModelsFitPrivacy-AwareCloudSharing?
• Ex: With which privacy level would you
share a project presentation with a friend?
• Standardized Infit Statistic:
• (x-axis values should lie in [-2,2])
16
81
96
Polytomous case
Infit t-statistic
Sensitivity
Dichotomous case
Sensitivity
Infit t-statistic
A dot
represents a
context](https://image.slidesharecdn.com/onlineversion-160710210747/85/C3P-Context-Aware-Crowdsourced-Cloud-Privacy-at-PETS-2014-54-320.jpg)
![IRTModelsFitPrivacy-AwareCloudSharing?
• Ex: With which privacy level would you
share a project presentation with a friend?
• Standardized Infit Statistic:
• (x-axis values should lie in [-2,2])
16
81
96
Polytomous case
Infit t-statistic
Sensitivity
Dichotomous case
Sensitivity
Infit t-statistic
A dot
represents a
context](https://image.slidesharecdn.com/onlineversion-160710210747/85/C3P-Context-Aware-Crowdsourced-Cloud-Privacy-at-PETS-2014-55-320.jpg)
![IRTModelsFitPrivacy-AwareCloudSharing?
• Ex: With which privacy level would you
share a project presentation with a friend?
• Standardized Infit Statistic:
• (x-axis values should lie in [-2,2])
16
81
96
Yes!
Polytomous case
Infit t-statistic
Sensitivity
Dichotomous case
Sensitivity
Infit t-statistic
A dot
represents a
context](https://image.slidesharecdn.com/onlineversion-160710210747/85/C3P-Context-Aware-Crowdsourced-Cloud-Privacy-at-PETS-2014-56-320.jpg)
Uploaded byHamza Harkous
C3P: Context-Aware Crowdsourced Cloud Privacy (at PETS 2014)
The document discusses C3P, a context-aware crowdsourced privacy solution for enhancing cloud data protection. It highlights the 60% increase in corporate data shared to the cloud in 2015, with 20% of those files containing protected information, and emphasizes the need for fine-grained data sensitivity assessment rather than brute encryption. It proposes a private crowdsourcing mechanism for estimating data sensitivity based on context and user input to improve privacy management.
More Related Content
Similar to C3P: Context-Aware Crowdsourced Cloud Privacy (at PETS 2014)
C3P: Context-Aware Crowdsourced Cloud Privacy (at PETS 2014)
- 1. C3P: Context-Aware Crowdsourced CloudPrivacy 1 CloudSpacesPrivacyEnhancingTechnologiesSymposium,2014
- 2.
- 3.
- 4.
- 5.
- 6.
- 7. 3 60% increase in corporatedata shared to the cloud in 2015 Source: Elastica’s Q2 2015 Shadow Data Report
- 8. 3 20% of files sharedto the cloud contain protected data 60% increase in corporate data shared to the cloud in 2015 Source: Elastica’s Q2 2015 Shadow Data Report
- 9. 3 20% of files sharedto the cloud contain protected data 60% of sensitive files contain PII 30% …contain health info 60% increase in corporate data shared to the cloud in 2015 Source: Elastica’s Q2 2015 Shadow Data Report
- 10. 3 20% of files sharedto the cloud contain protected data 60% of sensitive files contain PII 30% …contain health info Emergence of“Shadow IT” 60% increase in corporate data shared to the cloud in 2015 Source: Elastica’s Q2 2015 Shadow Data Report
- 11. You cannot use cloudservices. You are fully protected. Your files are always encrypted before uploading. Anti-Snooping Tools for the Cloud Examples: 4
- 12. You cannot run software. You arefully protected. Your files are always quarantined. WhatifAntivirusSoftwarewasSimilar? 5
- 13.
- 14. Obstacles Privacy vs. Servicesdilemma Context-dependence of privacy
- 15. Obstacles I dedicate therest of my life for sorting out sensitive from non-sensitive files on my HD Privacy vs. Services dilemma Context-dependence of privacy Manual effort and expertise for assessing data sensitivity 6
- 16. What is needed? Ensureserviceable protection instead of brute encryption.
- 17. What is needed? Ensureserviceable protection instead of brute encryption. Account for the metadata, sharing environment, and data content.
- 18. What is needed? Idedicate the rest of my life for sorting out sensitive from non-sensitive files on my HD Ensure serviceable protection instead of brute encryption. Account for the metadata, sharing environment, and data content. Automatically estimate the sensitivity of shared data. 7
- 19. Introducing C3P Various levelsof information hiding 8
- 20. Introducing C3P Define datain terms of context Various levels of information hiding 8
- 21. Introducing C3P I dedicatethe rest of my life for sorting out sensitive from non-sensitive files on my HD Private crowdsourcing mechanism for gathering people privacy policies Define data in terms of context Various levels of information hiding 8
- 22. Introducing C3P I dedicatethe rest of my life for sorting out sensitive from non-sensitive files on my HD Private crowdsourcing mechanism for gathering people privacy policies Psychologically grounded approach for estimating sensitivity Define data in terms of context Various levels of information hiding 8
- 23.
- 24. Defining Data throughContext 10 Content Metadata Environment
- 25. Defining Data throughContext 10 Content Metadata Environment
- 26. Defining Data throughContext 10 Content Metadata Environment
- 27.
- 28. PrivacyPreserving Crowdsourcing 12 Business MeColleagueFinancial Me Stranger Faces Home Friend Financial Me Stranger Business Me Colleague Faces Home Friend I dedicate the rest of my life for sorting out sensitive from non-sensitive files on my HD User 1 User 2 User 3
- 29. PrivacyPreserving Crowdsourcing 12 Business MeColleagueFinancial Me Stranger Faces Home Friend Financial Me Stranger Business Me Colleague Faces Home Friend Faces Home Friend Sharing Operation Context I dedicate the rest of my life for sorting out sensitive from non-sensitive files on my HD User 1 User 2 User 3
- 30. PrivacyPreserving Crowdsourcing 12 Business MeColleagueFinancial Me Stranger Faces Home Friend Financial Me Stranger Business Me Colleague Faces Home Friend Faces Home Friend Work Sea Colleague Family Sharing Operation Context I dedicate the rest of my life for sorting out sensitive from non-sensitive files on my HD User 1 User 2 User 3
- 31. PrivacyPreserving Crowdsourcing 12 Business MeColleagueFinancial Me Stranger Faces Home Friend Financial Me Stranger Business Me Colleague Faces Home Friend Faces Home Friend Work Sea Colleague Family Forward-Anonymity K-anonymity Sharing Operation Context I dedicate the rest of my life for sorting out sensitive from non-sensitive files on my HD User 1 User 2 User 3 Faces Home Friend Context
- 32.
- 33.
- 34.
- 35. High Privacy Attitude 75% SensitivityEstimationusingItemResponseTheory 13 Faces HomeFriend High Sensitivity 75%
- 36. High Privacy Attitude 75% SensitivityEstimationusingItemResponseTheory 13 Faces HomeFriend High Sensitivity 75%
- 37. High Privacy Attitude 75% SensitivityEstimationusingItemResponseTheory 13 Faces HomeFriend High Sensitivity 75%
- 38. High Privacy Attitude 75% SensitivityEstimationusingItemResponseTheory 13 Faces HomeFriend High Sensitivity 75% Group Invariance Faces Home Friend Faces Home Friend
- 39. High Privacy Attitude 75% SensitivityEstimationusingItemResponseTheory 13 Faces HomeFriend High Sensitivity 75% Group Invariance Faces Home Friend Faces Home Friend Item Invariance
- 40.
- 41.
- 42. Connectingthe Dots 14 Financial MeStranger Client Server Context Extraction ?
- 43. Connectingthe Dots 14 Financial MeStranger Client Server Context Extraction Sensitivity Request ?
- 44. Connectingthe Dots 14 Financial MeStranger Client Server Sensitivity Reply ?
- 45. Connectingthe Dots 14 Financial MeStranger Client Server Sensitivity Reply Policy Decision ?
- 46. Connectingthe Dots 14 Financial MeStranger Client Server Policy Decision Data Sharing ?
- 47. Connectingthe Dots 14 Financial MeStranger Client Server Crowdsourcing ?
- 48. Connectingthe Dots 14 Financial MeStranger Client Server Crowdsourcing ? Sensitivity Computation
- 49.
- 50.
- 51. IRTModelsFitPrivacy-AwareCloudSharing? • Ex: Withwhich privacy level would you share a project presentation with a friend? 16 81 96
- 52. IRTModelsFitPrivacy-AwareCloudSharing? • Ex: Withwhich privacy level would you share a project presentation with a friend? • Standardized Infit Statistic: • (x-axis values should lie in [-2,2]) 16 81 96 Dichotomous case Sensitivity Infit t-statistic A dot represents a context
- 53. IRTModelsFitPrivacy-AwareCloudSharing? • Ex: Withwhich privacy level would you share a project presentation with a friend? • Standardized Infit Statistic: • (x-axis values should lie in [-2,2]) 16 81 96 Dichotomous case Sensitivity Infit t-statistic A dot represents a context
- 54. IRTModelsFitPrivacy-AwareCloudSharing? • Ex: Withwhich privacy level would you share a project presentation with a friend? • Standardized Infit Statistic: • (x-axis values should lie in [-2,2]) 16 81 96 Polytomous case Infit t-statistic Sensitivity Dichotomous case Sensitivity Infit t-statistic A dot represents a context
- 55. IRTModelsFitPrivacy-AwareCloudSharing? • Ex: Withwhich privacy level would you share a project presentation with a friend? • Standardized Infit Statistic: • (x-axis values should lie in [-2,2]) 16 81 96 Polytomous case Infit t-statistic Sensitivity Dichotomous case Sensitivity Infit t-statistic A dot represents a context
- 56. IRTModelsFitPrivacy-AwareCloudSharing? • Ex: Withwhich privacy level would you share a project presentation with a friend? • Standardized Infit Statistic: • (x-axis values should lie in [-2,2]) 16 81 96 Yes! Polytomous case Infit t-statistic Sensitivity Dichotomous case Sensitivity Infit t-statistic A dot represents a context
- 57. TemporalCostofCrowdsourcing&Privacy 17 Zipf context distribution 500 3125 30000 av.:1 Item/6 hours • Synthetic Dataset:
- 58. TemporalCostofCrowdsourcing&Privacy k 17 Zipf context distribution 500 3125 30000 av.:1 Item/6 hours • Synthetic Dataset:
- 59. TemporalCostofCrowdsourcing&Privacy k 17 Zipf context distribution 500 3125 30000 av.:1 Item/6 hours • Synthetic Dataset: Crowdsourcing cost: Hit rate (HR) from 0 to 90% in 10 days
- 60. TemporalCostofCrowdsourcing&Privacy k 17 Zipf context distribution 500 3125 30000 av.:1 Item/6 hours • Synthetic Dataset: Crowdsourcing cost: Hit rate (HR) from 0 to 90% in 10 days Anonymity cost: HR difference starts high but vanishes in 10 days.
- 61. EffectofSharingBehaviorontheTemporalCost 18 Anonymity Parameter K=3 500 3125 30000 av.:1 Item/6 hours • Synthetic Dataset:
- 62. EffectofSharingBehaviorontheTemporalCost 18 Effect: Long taildistribution is of lower temporal cost. Anonymity Parameter K=3 500 3125 30000 av.: 1 Item/6 hours • Synthetic Dataset:
- 63. RobustnessTowardsMaliciousUsers? 19 • Test: • Assignsensitivities to items and attitudes to people. • Honest users choose policies according to the model. • Malicious users choose policies at random.
- 64. RobustnessTowardsMaliciousUsers? 19 • Test: • Assignsensitivities to items and attitudes to people. • Honest users choose policies according to the model. • Malicious users choose policies at random.
- 65. RobustnessTowardsMaliciousUsers? 19 • Test: • Assignsensitivities to items and attitudes to people. • Honest users choose policies according to the model. • Malicious users choose policies at random.
- 66. RobustnessTowardsMaliciousUsers? 19 • Test: • Assignsensitivities to items and attitudes to people. • Honest users choose policies according to the model. • Malicious users choose policies at random. Preset Sensitivity Computed Sensitivity-Check
- 67. RobustnessTowardsMaliciousUsers? 19 • Test: • Assignsensitivities to items and attitudes to people. • Honest users choose policies according to the model. • Malicious users choose policies at random. Preset Sensitivity Computed Sensitivity-Check Tolerance: 25% malicious: ≈8% difference, 50% malicious: ≈17% difference
- 68. FutureWork • Recommendation ofpolicies to users • Batch sensitivity analysis • Considering probabilistic attacks on the scheme • Working with IRT alternatives. 20 ELO MF
- 69.
- 70.
- 71. PrivyShareBenefits • Works withany cloud service 23
- 72. PrivyShareBenefits • Works withany cloud service • Provides“Sensitivity as a Service” 23
- 73. PrivyShareBenefits • Works withany cloud service • Provides“Sensitivity as a Service” • Offers fine grained protection • Metadata cleaning • Face Blurring • Encryption • Encryption + Auxiliary Information (automatic summaries, blurred thumbnails) 23
- 74.
- 75.
- 76.
- 77.