As a teenager circa 1995, I wanted to make my own guitar effects but struggled with the basics of electrical engineering. As a grad student in 2001, I tried again and built a Fuzz Face clone, but didn't have time to keep going. In 2014, I decided to try again and now I'm having a blast with it.

In 2012, I challenged myself to play a game every day of the year. I mostly succeeded.

I am also a big fan of wordplay, especially anagrams.
(Alas, adorable clownery is iffy as papal gag, mon ami!)

And I'm sort of getting into blogging and tweeting.


Upon moving to Pittsburgh, I joined the band delicious pastries. I play guitar, sing sometimes, and write songs. You can check out our records here:

You can find us on Spotify, Apple Music, Google Play, Amazon Music, SoundCloud, and Bandcamp!

Check out a few fan-posted videos from our live shows:

From 2009-2012, fellow machine learning researcher Jacob Eisenstein played bass with us (that's him wearing a striped shirt in the first video). The first record Pretty Please also features ML/NLP heavy-hitters Chris Dyer on cello and Nathan Schneider on violin.

In grad school at UW-Madison, I fronted a few bands, and released a solo record called Sketches in 2004. You can hear that record and assorted other tracks on SoundCloud. However, I have found being part of a band more enjoyable than playing solo or fronting a band myself. I also founded FAWM.ORG, an online community for musicians of all stripes, and annual songwriting challenge (to write 14 new songs each February). This is gradually becoming more of a research project for computational creativity tools and social network modeling.

Bayesians Against Discrimination

In the fall of 2009, I supported vector machines at the Pittsburgh G20 Protests (view a slideshow of our nerdy machine learning protest signs). At one point, I found myself marching behind The Daily Show's John Oliver:

For this, my left arm appears briefly on the October 1, 2009 episode (holding the "Bayesians Against Discrimination" sign at about 0:25).

Word Cloud

Since it seems all the rage, here is a summary of my research publications circa 2012 (courtesy of Wordle):



I enjoy teaching from time to time. Some of my past instructional activities:

Arbitrary Status Numbers

Because these things apparently matter.

I have an Erdős Number of 4:

  1. Paul ErdősRonald J. Gould: S. Burr, P. Erdős, R.J. Faudree, C.C. Rousseau, R.H Schelp, R.J. Gould, and M.S. Jacobson. Goodness of trees for generalized books. Graphs and Combinatorics 3(1): 1-6. 1987.
  2. Avrim Blum: A.L. Blum and R.J. Gould. Generalized degree sums and Hamiltonian graphs. ARS Combinatoria 35(A):35-54. 1993.
  3. Tom Mitchell: A.L. Blum and T.M. Mitchell. Combining labeled and unlabeled data with co-training. Proceedings of the Conference on Computational Learning Theory (COLT), pages 92-100. ACM Press, 1998.
  4. myself: A. Carlson, J. Betteridge, B. Kisiel, B. Settles, E.R. Hruschka Jr. and T.M. Mitchell. Toward an Architecture for Never-Ending Language Learning. In Proceedings of the Conference on Artificial Intelligence (AAAI), pages 1306-1313. AAAI Press, 2010.

This is particularly cool for me since I am "Burr S." and the lead author of paper #1 in the chain is "S. Burr." What's more, link #3 is an important paper in the literature for semi-supervised learning, a research area in which I've been somewhat active.

Also, I arguably have a Bacon Number of 3:

  1. Kevin BaconWilliam Devane: Hollow Man (2000)
  2. Leo Burmeister: Honky Tonk Freeway (1981)
  3. myself: Locust Grove (1992)

I like that my name is embedded in Mr. Bur-meister's. Admittedly, Locust Grove is an obscure docu-drama about George Rogers Clark that isn't even listed in IMDb. Furthermore, Burmeister went by a pseudonym (it was a non-union job: he played a peddler, I played GRC's nephew), but I still think it counts. I stopped pursuing a childhood acting career after that. Interestingly, I do have an IMDb page, but for a silly film I scored.

So in short, I have a Erdős-Bacon Number of 4 + 3 = 7. Not as impressive as Natalie Portman (6), but closer than Fred Alan Wolf (8), and tied with the likes of Stephen Hawking and Carl Sagan.

One of my life goals is to have a more properly-defined Erdős-Bacon Number of 6 or less.

Web Toys

Various webapps from over the years that use simple statistical NLP in fun ways.

Academic Genealogy

From American computer scientists to German theologians via Danish linguists!

Credit goes to Ray Mooney and Tim Wille for connecting the dots.

© 2024. All rights reserved ⇔ Lehrer's vilest drag.