## Posts Tagged ‘**differential privacy**’

## Tutorial Videos on and around Differential Privacy

http://dimacs.rutgers.edu/Workshops/DifferentialPrivacy/Slides/slides.html

The tutorial speakers covered connections between DP and a range of areas:

- Moritz Hardt: Differential private algorithms via learning theory
- Gerome Miklau: Query optimization techniques from the DB community
- Benjamin Pierce: Using PL techniques to automate and verify proofs of privacy
- Aaron Roth: Game-theoretic perspectives on privacy

- Slides (with corrections): http://www.cse.psu.edu/~asmith/talks/2012-08-21-crypto-tutorial.pdf
- Video (no corrections!): http://www.youtube.com/watch?v=dbWx62C5Q4o&list=PL3C6A9D61E40300E6&index=26

## DIMACS Workshop on Differential Privacy

Aaron Roth and I are running a 3 day interdisciplinary workshop on differential privacy at DIMACS (Rutgers), on October 24-26. This is immediately following FOCS, which is being held nearby, in downtown New Brunswick. The workshop will begin with a day of tutorials on differential privacy as understood in various communities (theory, databases, programming languages, and game theory), and will continue with two days of research talks and discussion.

Details of the workshop can be found here: http://dimacs.rutgers.edu/Workshops/DifferentialPrivacy/

(n.b.: some extra speakers have confirmed who are not yet on the web page).

As part of the program, we will also have a session of short (5-10 minute) talks from students, postdocs, and other interested parties. We encourage submission of abstracts for short talks. The solicitation is below.

See you all in October!

Aaron and Adam

**DIMACS Workshop on Differential Privacy across Computer Science**

October 24-26, 2012

(immediately after FOCS 2012)

Call for Abstracts — Short Presentations

The upcoming DIMACS workshop on differential privacy will feature invited talks by experts from a range of areas in computer science as well as short talks (5 to 10 minutes) by participants.

Participants interested in giving a short presentation should send an email to asmith+dimacs@psu.edu containing a proposed talk title, abstract, and the speaker’s name and affiliation. We will try to

accommodate as many speakers as possible, but

a) requests received before October 1 will get full consideration

b) priority will be given to junior researchers, so students and postdocs should indicate their status in the email.

More information about the workshop:

The last few years have seen an explosion of results concerning differential privacy across many distinct but overlapping communities in computer science: Theoretical Computer Science, Databases, Programming Languages, Machine Learning, Data Mining, Security, and Cryptography. Each of these different areas has different priorities and techniques, and despite very similar interests, motivations, and choice of problems, it has become difficult to keep track of this large literature across so many different venues. The purpose of this workshop is to bring researchers in differential privacy across all of these communities together under one roof to discuss recent results and synchronize our understanding of the field. The first day of the workshop will include tutorials, representing a broad cross-section of research across fields. The remaining days will be devoted to talks on the exciting recent results in differential privacy across communities, discussion and formation of interesting open problems, and directions for potential inter-community collaborations.

A tentative program and registration information can be found at

http://dimacs.rutgers.edu/Workshops/DifferentialPrivacy/

## Privacy in the NYT

An article in yesterday’s New York Times (front “page” on the web last night) does a good job of highlighting some of the intricacies of “privacy” in online social networks. The article links to a surprising number of technical research articles. There were also two quotes that stuck out.

‘“Technology has rendered the conventional definition of personally identifiable information obsolete,” said Maneesha Mithal, associate director of the Federal Trade Commission’s privacy division.’

This is not news to most computer scientists, but it is nice to hear it from the FTC. [On a related point, the FTC is holding the third of a series of roundtable discussions on electronic privacy today. Webcast here.]

The ending quote of the article, from Jon Kleinberg, was more of a downer:

“When you’re doing stuff online, you should behave as if you’re doing it in public — because increasingly, [you are].”

I disagree with the most literal interpretation of the quote, since there are still many ways to do things privately online. But keeping your privacy increasingly requires both technical sophistication and great care. And of course that endangers some of the coolest things about the Internet.

## IPAM Workshop Wrap-Up

Last week was the Statistical and Learning-Theoretic Challenges in Data Privacy, which I co-organized with Cynthia Dwork, Steve Fienberg and Sesa Slavkovic. As I explained in my initial post on the workshop, the goal was to tie together work on privacy in statistical databases with the theoretical foundations of learning and statistics.

- Slides for most talks are online
- Blog posts: Arvind N., Jon K. #1, #2 (see also an older post by Ben R.)

The workshop was a success. For one thing, I got a new result out of it and lots of ideas for problems to work on. I even had fun most of the time^{1}.

### — A shift in tone —

More importantly, I felt a different tone in the conversations and talks at this workshop than at a previous ones involving a similar crowd. For the first time, most participants seemed to agree on what the important issues are. I’ve spent lots of time hanging out with statisticians recently, so this feeling may not have been shared by everyone. But one change was objectively clear: the statisticians in the crowd have become much better at describing their problems in computational terms. I distinctly remember encountering fierce resistance, at the original 2005 CS-Stats privacy workshop in Bertinoro, when we reductionist CS types tried to get statisticians to spell out the procedures they use to analyze social science data.

“Analysis requires judgement. It is as much art as science,” they said (which we translated as, “Recursion, shmecursion. We do not know our own programs!”).

“But can’t you try to pin down some common objectives?”, we answered.

This week, there were algorithms and well-defined objectives galore. It helped that we had some polyglots, like Martin Wainwright and Larry Wasserman, around to translate.

### — The “computational lens” at work —

An interesting feature of several talks was the explicit role of “computational” perspective. Both Frank McSherry and Yuval Nardi used techniques from numerical analysis, namely gradient ascent and the Newton-Raphson method, to design protocols which were both more efficient and easier to analyze than previous attempts based on a more global, structural perspective. Frank described a differentially private algorithm for logistic regression, joint with Ollie Williams; Yuval described an efficient SFE protocol for linear regression, joint with Steve Fienberg, Rob Hall, and others.

### — Two under-investigated ideas —

At the wrap-up session (see the notes), I pointed out two directions that I think have been investigated with much less rigor than they deserve:

#### “Cryptanalysis” for database privacy

It would be nice to have a systematic study of, and standard nomenclature for, attacks on privacy/anonymity in statistical databases. Right now it seems every paper ends up defining (or not defining) a model from scratch, yet many papers are doing essentially the same thing in different domains. Even an incomplete taxonomy would be helpful. Here are a few terms I’d like to see becoming standard:

- linkage attack
- reconstruction attack
- composition attack (my personal favorite)

On a related point, it would be nice to see a good categorization of the kinds of side information that gets used. For example, Johannes Gehrke at Cornell and his students have a few papers laying out categories of side information (I have issues with some of the positive results in those papers, but I think the quantification of side information is interesting).

#### Relaxed definitions of privacy with meaningful semantics

This is probably a topic for a much longer post, but briefly: it would be nice to see meaningful definitions of privacy in statistical databases that exploit the adversary’s uncertainty about the data. The normal approach to this is to specify a set of allowable prior distributions on the data (from the adversary’s point of view). However, one has to be careful. The versions I have seen are quite brittle. Some properties to keep in mind when considering new definitions:

- Composition
- Side information: is the class of priors rich enough to incorporate complex side information, such as an anonymization of a related database? [see composition above]
- Convexity and post-processing, as in Dan Kifer’s talk
- Equivalent, “semantic” characterizations [e.g. here, here]

### — Other notes —

- The majority of the talks were completely or partly on differential privacy. Notable exceptions: Brad Malin, Xiaofeng Wang, Ravi Kumar, Jiashun Jin, Yuval Nardi. Our goal was not to have such a preponderance of differential privacy talks, but some of the people we expected to talk about other things (like Jerry Reiter) decided to focus on differential privacy. Tailoring the talk to the crowd?
- The nonspeaker participants were heavily skewed towards CS. In particular, at least [see comments!] four professors (Gerome Miklau, Anupam Gupta, Jonathan Katz, Yevgeniy Dodis) and three postdocs (Katrina Liggett, Anand Sarwate, Arvind Narayanan) from CS departments attended just to listen to the talks; I recognized only one stats postdoc (Saki Kinney). I also recognized lots of UCLA locals there too from CS (Yuval Ishai, Rafi Ostrovsky, Amit Sahai) but none from statistics.
- The rump session + posters combination worked very well (despite my earlier doubts). Rump session slides are online.

^{1}Serious sleep deprivation due to jet-lagged kids and talk prep made the “fun” part occasionally difficult.

## Differential privacy and the secrecy of the sample

(This post was laid out lazily, using Luca‘s lovely latex2wp.)

** — 1. Differential Privacy — **

Differential privacy is a definition of “privacy” for statistical databases. Roughly, a statistical database is one which is used to provide aggregate, large-scale information about a population, without leaking information specific to individuals. Think, for example, of the data from government surveys (*e.g.* the decennial census or epidemiological studies), or data about a company’s customers that it would like a consultant to analyze.

The idea behind the definition is that *users*–that is, people getting access to aggregate information–should not be able to tell if a given individual’s data has been changed.

More formally, a data set is just a subset of items in a domain . For a given data set , we think of the server holding the data as applying a randomized algorithm , producing a random variable (distributed over vectors, strings, charts, or whatever). We say two data sets are *neighbors* if they differ in one element, that is, .

Definition 1A randomized algorithm is -differentially private if, for all pairs of neighbor data sets , and for all events in the output space of :

This definition has the flavor of *indistinguishability* in cryptography: it states that the random variables and must have similar distributions. The difference with the normal cryptographic setting is that the distance measure is multiplicative rather than additive. This is important for the semantics of differential privacy—see this paper for a discussion.

I hope to write a sequence of posts on differential privacy, mostly discussing aspects that don’t appear in published papers or that I feel escaped attention.

** — 2. Sampling to Amplify Privacy — **

To kick it off, I’ll prove here an “amplification” lemma for differential privacy. It was used implicitly in the design of an efficient, private PAC learner for the PARITY class in a FOCS 2008 paper by Shiva Kasiviswanathan, Homin Lee, Kobbi Nissim, Sofya Raskhodnikova and myself. But I think it is of much more general usefulness.

Roughly it states that given a -differentially private algorithm, one can get an -differentially private algorithm at the cost of shrinking the size of the data set by a factor of .

Suppose is a -differentially private algorithm that expects data sets from a domain as input. Consider a new algorithm , which runs on a random subsample of points from its input:

Algorithm 2 (Algorithm )On input and a multi-set

- Construct a set by selecting each element of independently with probability .
- Return .

Lemma 3 (Amplification via sampling)If is -differentially private, then for any , is -differentially private.