I gave an invited talk on crypto voting. Ran out of time, as usual. I need to find a way to describe this stuff more efficiently. Later, I also presented MarkPledge2, joint work with Andy Neff.
Olivier de Marneffe presented the UCL/Helios implementation. So much material to present there, he did a fantastic job of focusing on the core parts.
Ari Feldman presented some work on surprising covert channels on the bulletin board. Seems like one of the good themes for the year.
Vanessa Teague delved into the details of Pret-a-Voter voting.
Missed a couple of talks after lunch as I got sucked into other conversations and work.
Eric Rescorla, the man who speaks faster than anyone I know, talked about auditing processes when you don’t have good random numbers (you know, the real world.)
Rebecca Wright talked about interstate registration database matching, something that seems easy but requires a bit more thinking than you’d initially guess, as most of these voting topics generally do.
Jennifer King described new, released software to help automate the auditing process. It’s crucial to build these tools, I think, so we can make sure to validate our theories and move the process of election auditing along.
Last, a particularly interesting (and apparently controversial) presentation by Joe Hall on risk-limiting audits, in other words, if I got this right, audit procedures that upper-bound the probability of being wrong.
All in all, a very successful couple of days, a very good group of people, and a beautiful job done by the program chairs.