[liberationtech] TOMORROW - June 12 - Differential Privacy in the Real World: Imperfect Randomness and Floating-Point Arithmetic - Ilya Mironov

Yosem Companys companys at stanford.edu
Mon Jun 11 18:02:02 PDT 2012


From: Ankur Taly <ataly at stanford.edu>

TITLE: Differential Privacy in the Real World: Imperfect Randomness and
Floating-Point Arithmetic

SPEAKER: Ilya Mironov <mironov at microsoft.com>

ABSTRACT:

Differential privacy, introduced by Dwork et al in 2006, has become a de
facto standard for definition of privacy for aggregate computations over
sensitive data. In this talk we consider two challenges in implementing
differential privacy that remarkably result in very similar procedures for
evaluating one basic differentially-private mechanism.

The first challenge is implementing differential privacy with access only
to a single imperfect source of randomness. It has been shown that
information-theoretic private-key encryption is impossible unless the
source of randomness is extractable (Bosley and Dodis, TCC'07). Somewhat
surprisingly, we show that the answer is reversed if the target notion of
security is differential privacy. We prove that an accurate and
differentially-private approximation to the Laplacian mechanism is feasible
even if the only source of randomness accessible to the mechanism is
semi-random (Santha-Vazirani).

Another problem area confronting implementations of differential privacy in
the real world is floating-point arithmetic. We demonstrate that all four
publicly available general-purpose systems for differential privacy are
vulnerable to the same attack exploiting the gap between mathematical
abstractions and their implementations. We also discuss a fix to the
problem, which coincides with the mechanism robust to imperfect randomness.

The talk is going to be self-contained, with all important concepts and
definitions introduced and developed as needed. The first part of the talk
is based on the joint work (to appear at CRYPTO'12) by Yevgeniy Dodis,
Adriana Lopez-Alt (both - NYU), and Salil Vadhan (Harvard, Stanford, and
MSR).


TIME and PLACE
June 12 2012 (Tuesday) at 1630 hrs
Gates 463A

_______________________________________________
liberationtech mailing list
liberationtech at lists.stanford.edu

Should you need to change your subscription options, please go to:

https://mailman.stanford.edu/mailman/listinfo/liberationtech

If you would like to receive a daily digest, click "yes" (once you click above) next to "would you like to receive list mail batched in a daily digest?"

You will need the user name and password you receive from the list moderator in monthly reminders. You may ask for a reminder here: https://mailman.stanford.edu/mailman/listinfo/liberationtech

Should you need immediate assistance, please contact the list moderator.

Please don't forget to follow us on http://twitter.com/#!/Liberationtech

----- End forwarded message -----
-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
______________________________________________________________
ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE





More information about the cypherpunks-legacy mailing list