photo

Jason Altschuler

Email: jasonalt [at] mit [dot] edu
Office: Stata D760

About

I am a fourth year PhD student at MIT. I am fortunate to be advised by Pablo Parrilo. My research is supported by an NSF GRFP grant.

Previously, I was an undergrad at Princeton (class of '16) under the wonderful supervision of Elad Hazan and Emmanuel Abbe. I spent 2 of my undergrad summers interning in tech (at Google/Google Research) and the other 2 summers interning on Wall Street (DE Shaw and Tower Research Capital).

In my free time, I like to play chess. I obtained an International Master norm in Spain in 2015.

Publications

Recent papers

  • Lyapunov Exponent of Rank One Matrices: Ergodic Formula and Inapproximability of the Optimal Distribution [pdf, slides]
    Jason Altschuler, Pablo Parrilo.

    • Journal version: SIAM Journal on Control and Optimization (SICON), 2020.
    • Conference version: Conference on Decision and Control (CDC), 2019.
  • Massively scalable Sinkhorn distances via the Nyström method [pdf]
    Jason Altschuler, Francis Bach, Alessandro Rudi, Jonathan Weed.
    Conference on Neural Information Processing Systems (NeurIPS), 2019.

  • Best arm identification for contaminated bandits [pdf]
    Jason Altschuler, Victor-Emmanuel Brunel, Alan Malek.
    Journal of Machine Learning Research (JMLR), 2019.

  • Online learning over a finite action set with limited switching [pdf, poster, talk, slides]
    Jason Altschuler, Kunal Talwar.

    • Journal version: Mathematics of Operations Research (MOR), 2020.
    • Conference version: Conference on Learning Theory (COLT), 2018.

  • Near-linear time approximation algorithms for optimal transport via Sinkhorn iteration [pdf, poster, NeurIPS short video]
    Jason Altschuler, Jonathan Weed, Philippe Rigollet.
    Conference on Neural Information Processing Systems (NeurIPS), 2017. (Selected for Spotlight presentation)

  • Inclusion of forbidden minors in random representable matroids [pdf]
    Jason Altschuler, Elizabeth Yang.
    Discrete Mathematics, 2017.

  • Greedy column subset selection: new bounds and distributed algorithms [pdf, poster, slides]
    Jason Altschuler, Aditya Bhaskara, Gang Fu, Vahab Mirrokni, Afshin Rostamizadeh, Morteza Zadimoghaddam.
    International Conference on Machine Learning (ICML), 2016.

High school

  • Rapid analysis and exploration of fluorescence microscopy images [link]
    Benjamin Pavie, Satwik Rajaram, Austin Ouyang, Jason Altschuler, Robert Steininger, Lani Wu, Steven Altschuler.
    Journal of Visual Experiments (JoVE), 2014.

Theses and other unpublished stuff

  • Approximation algorithms for Independent Set via semidefinite programming hierarchies and randomized rounding [pdf]
    Jason Altschuler, Matt Brennan.
    Expository final project for David Karger's Randomized Algorithms course, 2017.

  • Minimax rates for online learning with limited decision changes (superseded by this paper in MOR/COLT)
    Undergrad thesis for Princeton CS department, 2016. Advised by Elad Hazan.

  • Probabilistic linear Boolean classification
    Undergrad thesis for Princeton math department, 2016. Advised by Emmanuel Abbe.

Teaching

At MIT, I was a TA for the grad course 6.255/15.093 Optimization Methods in Fall 2019.

At Princeton, I was a TA for:
  • COS 511: Theoretical machine learning (grad course) -- Spring 2016
  • MAT 340: Applied algebra -- Fall 2015
  • MAT 216: Accelerated honors real analysis -- Fall 2014
  • MAT 217: Honors linear algebra -- Spring 2014
  • MAT 215: Honors real analysis -- Fall 2013