picture of me in Milan, Italy.

Daniel Alabi

alabid [at] g [dot] harvard [dot] edu

I am a Computer Science Ph.D. student at Harvard University, where I am advised by Salil Vadhan.

My research interests lie in theoretical computer science (algorithms & complexity theory) with a focus on optimization/learning with constraints. Example application domains include data privacy, algorithmic fairness, and machine learning.

Currently, my research is (mostly) supported by a Facebook Emerging Scholars Award. I spent the summer of 2018 at Microsoft Research New England mentored by Adam Kalai! In my first year at Harvard, under supervision by Cynthia Rudin and Margo Seltzer, I dabbled in interpretable ML models. Previously, I spent a year as a (non-PhD) graduate research scholar at Columbia University, working with Chris Wiggins and Eugene Wu. Before coming back to academia, I was a Software Engineer on the distributed systems team at MongoDB. I obtained my bachelor's degrees in Mathematics and Computer Science from Carleton College where I was a Kellogg International Scholar.

In my free (and not-so-free) time, I participate in ballroom activities as a member of Harvard's Competitive Ballroom Dance Team.

Research Papers

Unleashing Linear Optimizers for Group-Fair Learning and Optimization.
Daniel Alabi, Nicole Immorlica, Adam Tauman Kalai.
COLT, July, 2018  [arxiv]  [PMLR] 
Learning Certifiably Optimal Rule Lists for Categorical Data.
Elaine Angelino, Nicholas Larus-Stone, Daniel Alabi, Margo Seltzer, and Cynthia Rudin.
JMLR, June, 2018  [arxiv] 
Long Journal Version of KDD 2017, SysML 2018 papers with more analyses/results/experiments/proofs.
Systems Optimizations for Learning Certifiably Optimal Rule Lists.
Nicholas Larus-Stone, Elaine Angelino, Daniel Alabi, Margo Seltzer, Vassilios Kaxiras, Aditya Saligrama, and Cynthia Rudin.
SysML, February, 2018  [pdf] 
Learning Certifiably Optimal Rule Lists.
Elaine Angelino, Nicholas Larus-Stone, Daniel Alabi, Margo Seltzer, and Cynthia Rudin.
KDD, August, 2017  [pdf] 
PFunk-H: Approximate Query Processing using Perceptual Models.
Daniel Alabi and Eugene Wu.
Proceedings of the International Workshop on Human-in-the-Loop Data Analytics (HILDA), June, 2016  [pdf] 
Poster at North East Database Day, January, 2016.

Selected Projects (with Code Artifacts)

Collaborative LaTeX editor.
CORELS is a custom discrete optimization technique for building rule lists over a categorical feature space.
My research was on the use of human perceptual models to make interactive visualizations faster by leveraging approximation (with formal guarantees).
See my GitHub page for some more projects not listed here.


Check out some of my instructional worksheets.