Update on the NIST Temporal Map Challenge: Winners, Sprint #2, and Metrics!
7 views
Skip to first unread message
Christine Task
Jan 13, 2021, 5:00:50 PM1/13/21
to opendp-c...@g.harvard.edu, Howarth, Gary S. (Fed)
Hi all,
I wanted to give you a quick update on exciting recent happenings with the NIST Temporal Map Data Challenge. This Challenge focuses on differential privacy for temporal data (where one individual can contribute multiple records over time, dramatically increasing sensitivity for many queries) and geographic applications (where accuracy must be maintained across a wide array of heterogeneous map segments). This abstract problem underlies everything from epidemiology and disaster planning data, to advertising and location tracking.
I wanted to give you a quick update on exciting recent happenings with the NIST Temporal Map Data Challenge. This Challenge focuses on differential privacy for temporal data (where one individual can contribute multiple records over time, dramatically increasing sensitivity for many queries) and geographic applications (where accuracy must be maintained across a wide array of heterogeneous map segments). This abstract problem underlies everything from epidemiology and disaster planning data, to advertising and location tracking.
Sprint #1 focused on privatizing policing data for baltimore. It wrapped up just before the holidays, and if you'd like to see what happened, you can check out the detailed and illustrated Winner's Announcement here.
Sprint #2 just started! In this sprint we're working with multiple years of American Community Survey data. Come check it out, and maybe try your hand! There's no requirement to have participated in previous sprints to join in this one. A kick-off webinar will happen tomorrow, Thursday 1/14/21, at 12:30p ET, and you can register here. (The recording will also be available afterwards).
Finally, producing the privacy-preserving data isn't the only open research problem to address, there's also the question of how we decide utility. We just concluded our sister challenge focused on designing utility metrics, "A Better Meterstick for Differential Privacy", and the submissions are now available for public voting. There's some very interesting ideas here, and we'll be considering these in our scoring design for Sprint #3 of the DP challenge. Come take a look yourself and cast a vote for your favorite!
Thank you all for joining us on this adventure--we've got further interesting things ahead of us. I'll report back again when it's time to kick-off Sprint #3, but I hope you'll consider hopping over to the DrivenData site and following along with us there for Sprint #2.
--Christine Task
Reply all
Reply to author
Forward
0 new messages