A Privacy-Protecting Architecture for Collaborative Filtering via - - PowerPoint PPT Presentation
A Privacy-Protecting Architecture for Collaborative Filtering via Forgery and Suppression of Ratings Javier Parra-Arnau , David Rebollo-Monedero and Jordi Forn http://sites.google.com/site/javierparraarnau/ Department of Telematics
Leuven, Belgium September 15, 2011
1
2
The amount of information on the Web has grown exponentially since
3
A recommendation system is a filtering system that suggest
4
mation overl ion overload
ion overload ion overload tion overloa ion overload ion overload formation o ion overload ion overload ion overload
ion overload ion overload rmation ove
rmation ove Information
5
Users need to communicate their preferences to the recommender
80 76 71 71 67 62 54 51 38 34 25 25 16 12 7 7 7 3 3
Drama Thriller Comedy Action Adventure Crime Romance Sci-Fi War Mistery Documentary Animation Fantasy Horror Children Musical Western Film-Noir IMAX
6
The privacy risks perceived by users include computers “figuring
… … predictions Recommendation System
7
Submitting false information and refusing to give private information
Our approach relies upon the forgery and suppression of ratings
Recommendation System
… … predictions
the user has read these books
8
Our architecture protects user privacy to a certain extent
9
Mathematical formulation of the optimal trade-off among privacy,
Our proposal could be used in combination with other existing
r;s ri>0; P ri=½ qi>si>0; P si=¾
The state-of-the-art approaches may be classified according to these
10
3.2 + 1.5, 2.9 – 0.7, 4.1, 4.4 – 2.7 5.6, 3.3 + 1.0, 1.1, 3.4 – 0.1
recommendation system
The state-of-the-art approaches may be classified according to these
10
The state-of-the-art approaches may be classified according to these
10
central server ratings
11
Profiling is accomplished on the basis of user ratings Information items are classified as known or unknown Users may wish to submit ratings to unknown items (forgery) and
Recommendation System
unknown items known items
12
[Toubiana 10, Fredrikson 11] suggest representing user profiles as
We model the profile of a user as a probability mass function (PMF)
Movielens
80 76 71 71 67 62 54 51 38 34 25 25 16 12 7 7 7 3 3
Drama Thriller Comedy Action Adventure Crime Romance Sci-Fi War Mistery Documentary Animation Fantasy Horror Children Musical Western Film-Noir IMAX
Witty
Buddies
Clever
Fall in Love
Couple
Good Best Friends Offbeat
Emotional Slow Teenage Life Sincere
Human Spirit Human Nature
Jinni
Our architecture requires to estimate the actual profile of a user to help
13
books \ literature & fiction \ genre fiction
Passive attacker capable of crawling through the items rated by a user The attacker observes the apparent user profile t, a perturbed
14
Recommender
predictions ratings … …
… …
… …
forgery and suppression
We measure privacy as the Shannon entropy of the user’s apparent
Accordingly, privacy is compromised whenever the user’s preferences
15
n
i=1
number of categories
1 2 3 4
1 2 3 4
16
Category Extractor Communication Manager
Information Provider Recommendation System
Known / Unknown Items Classifier User Profile Constructor
x2
uncategorized item categorized item unknown item rated item known item
...
Forgery Alarm Suppression Alarm
Forgery and Suppression Generator
User side Network side
! !
17
Category Extractor
Information Provider Recommendation System
Known / Unknown Items Classifier User Profile Constructor
x2
uncategorized item categorized item unknown item rated item known item
...
Forgery Alarm Suppression Alarm
Forgery and Suppression Generator
User side Network side
! !
Block Functionality
with the recommender
information about the items explored by the user
Communication Manager
Description - Starting at the beginning, the book explores how JavaScript originated and evolved into
what it is today. A detailed discussion of the components that make up a JavaScript implementation follows, with specific focus on standards such as ECMAScript and the Document Object Model (DOM).
Category - books \ computers & internet \ web development Average Customer Review 4.5/5 Description - Stephen Hawking, one of the most brilliant theoretical physicists in history, wrote the
modern classic A Brief History of Time to help nonscientists understand the questions being asked by scientists today.
Category - books \ science Average Customer Review 4/5 Description - Written by soccer great and championship Stanford coach Bobby Clark, this book tells
you how, starting at point zero, an uninitiated coach can meld kids into a team and help them enjoy one of the most rewarding experiences of their youth.
Category - books \ sports \ coaching \ soccer Average Customer Review 4.5/5 Description - You’ve made it! Your baby has turned one! Now the real fun begins. From temper
tantrums to toilet training, raising a toddler brings its own set of challenges and questions — and Toddler 411 has the answers.
Category - books \ parenting & families \ parenting Average Customer Review 3/5
18
Category Extractor
Information Provider Recommendation System
Known / Unknown Items Classifier User Profile Constructor
x2
uncategorized item categorized item unknown item rated item known item
...
Forgery Alarm Suppression Alarm
Forgery and Suppression Generator
User side Network side
! !
Block Functionality
associated with the items downloaded by the Communication Manager
Communication Manager
19
Category Extractor
Information Provider Recommendation System
Known / Unknown Items Classifier User Profile Constructor
x2
uncategorized item categorized item unknown item rated item known item
...
Forgery Alarm Suppression Alarm
Forgery and Suppression Generator
User side Network side
! !
Block Functionality
the items as known
Communication Manager
unknown items known items
books \ computers & internet \ web development books \ science books \ sports \ coaching \ soccer books \ parenting & families \ parenting
20
Category Extractor
Information Provider Recommendation System
Known / Unknown Items Classifier User Profile Constructor
x2
uncategorized item categorized item unknown item rated item known item
...
Forgery Alarm Suppression Alarm
Forgery and Suppression Generator
User side Network side
! !
Block Functionality
actual user profile
Communication Manager
… …
21
Category Extractor
Information Provider Recommendation System
Known / Unknown Items Classifier User Profile Constructor
x2
uncategorized item categorized item unknown item rated item known item
...
Forgery Alarm Suppression Alarm
Forgery and Suppression Generator
User side Network side
! !
Block Functionality
architecture
forgery rate ½ and a suppression rate ¾
Communication Manager
5% 8% 2%
¾ = 10% ½ = 5%
FORGERY SUPPRESSION
22
Category Extractor
Information Provider Recommendation System
Known / Unknown Items Classifier User Profile Constructor
x2
uncategorized item categorized item unknown item rated item known item
...
Forgery Alarm Suppression Alarm
Forgery and Suppression Generator
User side Network side
! !
Block Functionality
when an item should be suppressed
Communication Manager
science parenting
8% 2%
23
Category Extractor
Information Provider Recommendation System
Known / Unknown Items Classifier User Profile Constructor
x2
uncategorized item categorized item unknown item rated item known item
...
Forgery Alarm Suppression Alarm
Forgery and Suppression Generator
User side Network side
! !
Block Functionality
when an item should be forged
Communication Manager
computers sports population’s rating
5%
computers
The degradation in the accuracy of predictions is measured as ¾ and ½ We model items as r.v.’s taking on values in a common finite alphabet
We define
Accordingly, the user’s apparent profile is defined as
24
½ 2 [0; 1) ¾ 2 [0; 1)
Privacy is measured as the Shannon entropy of the user’s apparent
The privacy-forgery-suppression function
25
r;s ri>0; P ri=½ qi>si>0; P si=¾
Forgery and Suppression Generator
The forgery and suppression of ratings arise as two simple
We propose an architecture that implements these two mechanisms in
We present a formulation of the optimal trade-off among privacy,
26
Leuven, Belgium September 15, 2011
39