Workshop on Design Issues for a data Anonymization Competition (WODIAC)

The analysis of large scale datasets, often refer to as Big Data, offers the possibility to realize inferences with an unprecedented level of accuracy and details. However, this massive collection of information also raises many privacy issues since most of these datasets contain personal information, which is thus sensitive by nature. As a result, only very few of them are released and available, which limits both our ability to analyze such data to derive useful knowledge that could benefit to the public and the society at large and slows down the innovative services that could emerge from such data. Thus, an important scientific and societal challenge is the design and study anonymization mechanisms that can be used to remove the sensitive information or add uncertainty to a dataset before it is released or before further services are developed on it.

To address this issue, we will organize a data anonymization competition, to be held in conjunction with PETS 2018, to study the strengths and limits of anonymization methods from an empirical perspective. The main objective of this workshop is to investigate the design issues related to the organization of such a competition:

• First, the choice of the privacy model, the privacy metrics as well as the adversary model used are fundamental issues that need to be well thought out to be sure that they really correspond to realistic and grounded measures of privacy.
• Second, it is also necessary to define the main inference attacks that could be performed by the adversary based on the released data but also on his background knowledge.
  
• Third, the utility measures need also be chosen to meaningfully assess the achievable trade-off between utility and privacy.
  
• Fourth, the logistical issues surrounding the competition such as data sources, setting, rules, format and timeline.
  

The expected outcome of the workshop is a set of recommendations and guidelines to establish the competition.

Organizers

Sébastien Gambs, UQAM (Université du Québec à Montréal), Canada, <gambs DOT sebastien AT uqam ca>
Hiroaki Kikuchi, Meiji University, Japan, <kikn AT meiji ac jp>

Call for Contributed Talks

If you are interested to give a talk that fits within the scope of the workshop please send us a title and a one or two paragraphs description of our proposed presentation before June 25 to both of the chairs at the email addresses listed above. We are planning to release the detailed program of the workshop by the end of June. There will be no official proceedings but the expected outcome of the workshop will a set of recommendations and guidelines to establish the competition based on the presentations and exchanges generated during the workshop.

Important Dates:

• Talk submission deadline: June 25
• Notification and tentative program: June 30
• Workshop: July 17

Registration

The workshop will be held from 8am-4pm on Monday, 17 July 2017 on the University of Minnesota campus.  All participants are required to register, via learning.umn.edu.  Registration is $35 and covers printed materials, a morning coffee break, and box lunch.  Workshop participants that are not planning to attend PETS can also register to stay in a single-person room at the University of Minnesota Dormitory for two nights, July 15-16, at a cost of $56.40/night.

Program

Monday, 17 July, 2017

Keller Hall 3-180

8:30 - 9:00 Registration and morning coffee (Keller Hall 3-176)

9:00 - 9:15 Welcome and opening

9:15 - 10:45 Session 1: Feedback from the Japanese anonymization competition

• Study from Data Anonymization Competition of Online Retail Data
  Hiroaki Kikuchi (Meiji University)
• Quantifying the Risk of Re-identification in Data Anonymization Competition
  Takao Murakami (AIST)
• Anonymization and Re-identification for Personal Transaction Data
  Hiroshi Nakagawa (University of Tokyo/RIKEN AIP)

10:45 - 11:00 Coffee Break (Keller Hall 3-176)

11:00 - 12:30 Session 2: Location privacy

• Importance of realistic adversary model for anonymity evaluation: A case study of trajectory data
  Shogo Masaki (NTT Corp.)
• Discrimination Rate: An Attribute-Centric Metric to Measure Privacy
  Louis-Philippe SONDECK (Orange/IMT/OLPS/ASE/IDEA/PIA)
• Inference attacks on location data
  Sébastien Gambs, UQAM (Université du Québec à Montréal), Canada
  

12:30 - 14:00 Lunch Break (Keller Hall 3-176 and atrium)

14:00 - 16:00 Session 3: Panel discussions (slides)

16:00- 16:15 Closing