Apple’s ‘Differential Privacy’ Is About Collecting Your Data—But Not Your Data

ronjor · Jun 14, 2016

Andy Greenberg 06.13.16.
Apple, like practically every mega-corporation, wants to know as much as possible about its customers. But it’s also marketed itself as Silicon Valley’s privacy champion, one that—unlike so many of its advertising-driven competitors—wants to know as little as possible about you. So perhaps it’s no surprise that the company has now publicly boasted about its work in an obscure branch of mathematics that deals with exactly that paradox.
Click to expand...

https://www.wired.com/2016/06/apples-differential-privacy-collecting-data/

lotuseclat79 · Jul 7, 2016

What is Differential Privacy?

...Apple announced a series of new security and privacy features, including one feature that's drawn a bit of attention -- and confusion. Specifically, Apple announced that they will be using a technique called "Differential Privacy" (henceforth: DP) to improve the privacy of their data collection practices.
Click to expand...

-- Tom

mirimir · Jul 8, 2016

It sounds good. But how would outsiders ever know what they do?

lotuseclat79 · Aug 26, 2016

Differential Privacy is Vulnerable to Correlated Data — Introducing Dependent Differential Privacy

To provide its guarantees, DP implicitly assumes that the data tuples in the database, each from a different user, are all independent. This is a weak assumption, especially because tuple dependence occurs naturally in datasets due to social interactions between users. For example, consider a social network graph with nodes representing users, and edges representing friendship relationships. An adversary can infer the existence of a social link between two users that are not explicitly connected in the graph using the existence of edges among other users in the graph ( since two friends are likely to have many other friends in common). Similarly, private attributes in a user’s record can be inferred by exploiting the public attributes of other users sharing similar interests. An adversary can easily infer a user’s susceptibility to a contagious disease using access to the noisy query result and also the fact that the user’s immediate family members are part of the database. In an era of big-data, the tuples within a database present rich semantics and complex structures that inevitably lead to data correlation. Therefore, data correlation, i.e., the probabilistic dependence relationship among tuples in the database, which is implicitly ignored in DP, can seriously deteriorate the privacy properties of DP.
Click to expand...

-- Tom

Log in or Sign up

Apple’s ‘Differential Privacy’ Is About Collecting Your Data—But Not Your Data

ronjor Global Moderator

lotuseclat79 Registered Member

mirimir Registered Member

lotuseclat79 Registered Member

Log in or Sign up

Apple’s ‘Differential Privacy’ Is About Collecting Your Data—But Not Your Data

ronjor Global Moderator

lotuseclat79 Registered Member

mirimir Registered Member

lotuseclat79 Registered Member

Useful Searches