Sébastien Bubeck

Portrait 
Senior Researcher

Machine Learning and Optimization, Microsoft Research, Redmond

Contact

Building 99, 2938

Redmond, WA 98052

sebubeck AT microsoft DOT com

Co-chair for COLT 2018.

Associate Editor for Mathematics of Operations Research.

Associate Editor for Mathematical Statistics and Learning (publisher: European Mathematical Society)

Steering committee member (elected) for COLT from 2014 to 2017.

I was co-general chair for COLT 2013, COLT 2014, and I was on the program committee for NIPS 2012, NIPS 2014, NIPS 2016, NIPS 2017, COLT 2013, COLT 2014, COLT 2015, COLT 2016, COLT 2017, ICML 2015, ICML 2016, ICML 2017, SODA 2017, Random 2017, ALT 2013, ALT 2014.

In addition to my longstanding interest in online decision making under uncertainty I am also currently interested in (i) the interplay between convexity and randomness in optimization, and (ii) inference problems on random graphs.

NEW: A regularization approach for k-server and metrical task systems with the multiscale entropy

With a fantastic team of co-authors (Michael Cohen, James R. Lee, Yin Tat Lee, Aleksander Madry) we improved the state of the art competitive ratio for k-server and metrical task systems by using the mirror descent algorithm. To learn more about it I recommend to first take a look at this [youtube video], then these 3 blog posts ([part 1], [part 2], [part 3]) and finally [the k-server paper] itself (the MTS paper is coming up soon).

Polynomial-time algorithm for bandit convex optimization

From July 2014 to July 2016 with various co-authors at MSR we dedicated a lot of energy to bandit convex optimization. The end product is a new efficient algorithm. To learn more about it I recommend to first take a look at this [youtube video], then these 3 blog posts ([part 1], [part 2], [part 3]), and finally [the paper] itself.

Research Interests

  • machine learning

  • convex optimization

  • multi-armed bandits

  • statistical network analysis

  • random graphs and random matrices

  • online algorithms (in particular metrical task systems and k-server)

  • applications of information theory to learning, optimization, and probability