A Privacy-preserving Collaborative Filtering Scheme with Two-way Communication

Source Type: Research paper
Source URL: http://portal.acm.org/ft_gateway.cfm?id=1134742&type=pdf&coll=GUIDE&dl=GUIDE&CFID=20955699&CFTOKEN=40773225

Summary: In traditional CF systems, a server first collects ratings from users and then executes CF algorithms to make recommendation. There is a serious threat to individual privacy since data collected from users cover personal information about places and things they do, watch, and purchases. To solve this issue, a randomization approach has been proposed to disguise user ratings while still producing accurate recommendations. However, recent research work[1] has point out that randomization techniques might not preserve privacy as much as had been believed. This paper introduce a two-way communication privacy-preserving scheme in which users perturb their ratings for each item based on the server's guidance instead of using an item-invariant perturbation. According to their experiment, their new scheme preserve more privacy information than the randomization approach at the same accuracy level.

Reference:
1. Deriving Private Information from Randomized Data