Targeted Advertising… and Privacy Too e! Buy m Ari Juels RSA Laboratories
Targeted advertising: One View e! Buy m Antidandruff Buy me to o Anti-wr ! inkle Cream
An example Special Report: October 5 th is America's most popular birthday.
Anybirthday. com targeted gift recommendations Hiking/sports book Anti-aging cream
Targeted advertising: The benign view What shall I read next? Consumer Here are some novels you might enjoy Anna Karenina Wild Things Red Chamber Bleak House Advertiser
The other view: Consumer privacy Browser cookies Consumer
Our Goal + Consumer Privacy Advertiser Efficiency
Private Profile Physical characteristics: Age: 30 Weight: 120 Height: 5’ 4” …. Favorite foods: Chocolate Sushi Broccoli Hobbies: Butterfly collecting Tai Chi Archery Financial (from Quicken): $110, 000 / year income House worth $300, 000 $20, 000 butterfly collection Web sites recently visited (from browser): aclu. org yahoo. com anybirthday. com
Negotiant Advertiser Profile Negotiant Ad request
Negotiant Alice if annual income > $100, 000 if likes funny clothes or CEO output “golf ad” else if female output “spa ad” else output “cigar ad” else output “Walmart ad” Advertiser “spa ad”
Private Information Retrieval (PIR) Spa ad Advertiser’s server Alice’s computer Negotiant “spa ad” Problem: Expensive! Ads
Scheme 1: Naïve PIR Spa ad Ads Advertiser’s server Alice’s computer Negotiant “spa ad” Ads Problems: Inefficient; no idea what ads were distributed
Scheme 2: Direct request Spa ad “spa ad” please Spa ad Advertiser’s server Alice’s computer Negotiant “spa ad” Ads
Basic tool: Mix network plaintext 1 Mix network plaintext 2 plaintext 3 plaintext 4 Randomly permutes and encrypts inputs
Opposite direction Mix network plaintext 1 plaintext 2 plaintext 3 plaintext 4 Randomly permutes and decrypts inputs
Scheme 3: Semi-private PIR “spa Alice’s computer ad” “Walmart ad” “golf ad” Bob’s computer ad” art lm “Wa t ar m l Carol’s computer a “W Darius’s computer ” ad “Walmart ad” Advertiser’s server “spa ad” Mix network Ads
Scheme 3: Semi-private PIR Spa ad spa Alice’s computer ad Walmart ad golf ad Walmart ad Bob’s computer d rt a a m Wal Carol’s computer Darius’s computer t a W ar m l ad Advertiser’s server spa ad Mix network Ads
Scheme 3: A little more detail Alice, t ques ad re name , “spa ad”
Scheme 3: A little more detail Alice, ad name Spa ad , , “spa ad”
Remarks u We assume advertiser may collude with some mix servers (if not, one-server mix will do) u Ads are long, so we need a hybrid mix network – New constructions: OA 00, JJ 01 u Other tools – Quorum controlled threshold proxy reencryption u Aggregation and mixing can be offline
Malicious negotiant Negotiant Advertiser Profile ad number = annual income “ad $110, 000” please
Spotting Malicious Negotiants u Open source negotiants u “Sandbox” approach u Third-party validation
Scheme 4: Threshold PIR u Full set of ads mixed u Full privacy u Plaintext Equality Testing (PET) primitive needed u High cost (still better than PIR)
What are the (hoped for) benefits? u More consumer trust – Higher “opt in” rate – Truth in (requested) advertising u Complementary with P 3 P u Mobility possible through encrypted profile on server
Questions?
Скачать презентацию Targeted Advertising and Privacy Too e Buy m
cab7e995c2f2758d163bdff1bd212332.ppt