Author Archives: Anand Sarwate

Jácaras and Ensaladas

Posted on by

I am singing with the UChicago Early Music Ensemble, a somewhat relaxed group led by David Douglass and Ellen Hargis of The Newbury Consort. I started rehearsing a bit late, so I’ve been playing catch-up. This year the repertoire is all music from Spain and Spanish colonies, and today we worked on two of the harder pieces in the program : an ensalada called La Bomba, by Mateo Flecha “El Viejo”, and a jacara by Juan Gutiérrez de Padilla. Too many words!

In looking for some recordings to get a better sense of the pieces, I came across this charmingly old King’s Singers TV special (check out those sweater vests!) acting out La Bomba in what appears to be the house from Clue:

Lucky them, they get multiple takes which makes it a bit easier to manage the crazy transitions in the piece. There’s also a multitracked recording on which is pretty good:

Unfortunately we are doing it up a third from there, much to the chagrin of my passagio.

We spent a bit of time trying to get the jácara up to speed, but singing Spanish that fast is hard! When I heard how fast this version went I almost lost it:

It looks like I have my work cut out for me, especially if I want to roll my r’s like that.

New books on the foundations of signal processing

Posted on by

For those who are interested in signal processing, the classic Wavelets and Subband Coding by Martin Vetterli and Jelena Kovačević is a must-have. But perhaps you were unaware that the two of them, together with Vivek Goyal have written a pair of books:

Signal Processing: Foundations
Signal Processing: Fourier and Wavelet Representations

They should give you a solid grounding the fundamentals of signal processing from a more modern perspective. Preliminary versions are available for download from the book’s website.

Home/Land at Albany Park Theatre Project

Posted on by

On Friday I saw Home/Land at the Albany Park Theatre Project. I’ve been raving to people about it because I think this may be the most important piece of theater I have ever seen.

This is a play built out of stories collected from all over Chicago about immigration and the struggles of families and communities who have come here from all over the world. The actors are high schoolers, the youth of those communities and they bring with them an urgency that is palpable. These stories need to be told, and it is precisely that need that transforms the theater for those two hours. The interior imaginations of these performers is rich, surprising, and incisive. This is a kind of total theater — physical movement, song, and ritual — that you would not expect to come from, well, kids. And you wouldn’t expect to see it in the commercial theater.

The first clip in this profile of the show on PBS was one of the most moving moments in the piece — when I saw it each beat came perfectly timed, the choreographed raw anguish casually brushed away by two guards righting the door. These are real things that happen to real people, and we too can brush these things off. As the saying goes, “attention must be paid.”

The show has been extended but is sold out. Get on the wait list. They will call you if there is space, and if you get in, you will not regret it. If you don’t believe me, read some other more professional reviewers.

Truth in surveying

Posted on by

A few weeks ago I attended Scott Kominers‘s class on Market Design. They were talking about mechanism design and differential privacy so I felt like it would be fun to attend that session. In the class Scott mentioned some interesting work by Nicholas Lambert and Yoav Shoham on Truthful Surveys that appeared at WINE 2008. There’s also some recent work by Aaron Roth and Grant Schoenebeck up on ArXiV.

In Lambert and Shoham’s set up, the opinion distribution of a population is given by some CDF F(x) (with a density) on the unit interval [0,1]. We can think of x as a level of approval (say of a politician) and F(x) as the proportion of the population which has approval less than x. A surveyor selects n agents \{x_i\} i.i.d. from F and asks them to report their opinion. They can report anything they like, however, so they will report \{r_i\}. In order to incentivize them, the surveyor will issue a payment \Pi_i( r_1, \ldots, r_n ) to each agent i. How should we structure the payments to incentivize truthful reporting? In particular, can we make a mechanism in which being truthful is a Nash equilibrium (“accurate”) or the only Nash equilibrium (“strongly accurate”)?

Let A_i = |\{j : r_i r_j \}|. They propose partitioning the agents into k groups with \mathcal{G}(i) denoting the group of agent $i$, and \tilde{F}_i(x) as an unbiased estimator of F(x) that uses the points \{r_j : \mathcal{G}_j \ne \mathcal{G}_i \}. The payments are:

\Pi_i(\{r_j\}) = \frac{1}{|\mathcal{G}_i| - 1} \left[ A_i - B_i \right] + 2 \tilde{F}_i(r_i) - \frac{2}{|\mathcal{G}_i| - 1} \sum_{j \in \mathcal{G}_i \setminus \{i\} } \tilde{F}_j(r_j)

This mechanism is accurate and also permutation-invariant with respect to the agents (“anonymous”) and the sum of the payments is 0 (“budget-balanced”).

This is an instance of a more general mechanism for truthfully inducing samples from a collection of distributions that are known — each agent has a distribution F_i and you want to get their sample of that distribution. Here what they do is replace the known distributions with empirical estimates, in a sense. Why is this only accurate and not strongly accurate? It is possible that the agents could collude and pick a different common distribution G and report values from that. Essentially, each group has an incentive to report from the same distribution and then globally the optimal thing is for all the groups to report from the same distribution, but that distribution need not be F if there is global collusion. How do we get around this issue? If there is a set of “trusted” agents \mathcal{T}, then the estimators in the payment model can be built using the trusted data and the remaining untrusted agents can be put in a single group whose optimal strategy is now to follow the trusted agents. That mechanism is strongly accurate. In a sense the trusted agents cause the population to “gel” under this payment strategy.

It seems that Roth and Schoenbeck are not aware of Lambert and Shoham’s work, or it is sufficiently unrelated (they certainly don’t cite it). They also look at truth in surveying from a mechanism design perspective. Their model is somewhat more involved (an has Bayesian bits), but may be of interest to readers who like auction design.

Linkage

Posted on by

Congratulations to my fellow Beast Amitha Knight on being a co-winner of the 2012 PEN New Enlgand Susan P. Bloom Children’s Book Discovery Award!

Speaking of children’s books, some people who saw The Hunger Games movie are upset that Rue is black. Unsurprising but sad.

And speaking of friends, my friend Amber is slumming it in Antarctica and is writing some fascinating blog posts from down there.

Can Ellen Do More Push-Ups Than Michelle Obama? They both seem to be able to do more pushups than me. Time to hit the gym I think.

I’ve been eating this spicy peanut noodle salad for lunch this week and boy is it delicious.