Great Ideas in Theoretical Computer Science Seminar

Theoretical Computer Science is full of great ideas, so once a week, the MIT Theory group gets together to socialize and share some knowledge. We'll briefly share some interesting ideas, and discuss a superset of these ideas over lunch.

This webpage is for the 2015-2016 academic year.

Where: G5 Lounge, Stata Center
When: Usually Tuesday, Wednesday, or Thursday around lunch time. For exact dates and times:

Organizers: Rio LaVigne (rio@mit.edu) and Srinivasan Raghuraman (srirag@mit.edu).

  • Please contact us if you are want to give a talk or have any feedback.

Speaker Guidelines: If you are about to give a talk, please check out these guidelines.


Upcoming Talks

Time: Thursday, October 6, 12:15pm
Speaker: G Kamath
Title: Poisson *nomial Distributions

Time: Thursday, October 13, 12:15pm
Speaker: TBA
Title: TBA

Time: Thursday, October 20, 12:15pm
Speaker: TBA
Title: TBA


Past Talks

Time: Thursday, October 6, 12:15pm
Speaker: G Kamath
Title: Poisson *nomial Distributions
Abstract:
I'll talk about Poisson Binomial and Multinomial Distributions. While these are complex distributions with a large number of parameters, they enjoy remarkably elegant structural properties. I'll discuss these structural properties and applications to learning, game theory, elephants, limit theorems, and more.
Some of the results I present are based on joint work with Costis Daskalakis, Anindya De, and Christos Tzamos.



Time: Thursday, September 29, 12:15pm
Speaker: Michael Coulombe
Title: Investigating the Most Powerful Known Parallel Computer
Abstract:
The High Throughput Connectomics project aims to build a software pipeline based on theoretically sound principles and fast parallel algorithms to analyze brain samples and study the connectome: a "comprehensive map of neural connections in the brain." At a high level, I will discuss my work on the project in segmenting 3D blocks of images to identify the neurons passing through the volume and a surprising algorithm that I found which helped me beat the state-of-the-art.



Time: Thursday, September 22, 12:15pm
Speaker: Ilya Razenshteyn
Title: Transitive Closure of a Directed Graph in near-linear time
Abstract:
I will show how to estimate the size of the transitive closure of a directed graph in near-linear time using sampling. The algorithm is by Edith Cohen.