COR-OPT Seminar Spring 2018


Meeting Location and Time

Mondays 2-3PM in Rhodes Hall 253

Theme and Aims

The theme of this semester is “Statistics + Optimization.”

The aim of the seminar is to survey a recent thrust of work, occurring at the intersection of optimization and data science, which seeks to answer the following question:

When do statistical assumptions on input data lead to fast algorithms?

Key examples of this phenomenon: compressive sensing, matrix completion, phase retrieval, robust PCA, etc.


Anyone can volunteer to give a talk on their own research, provided it adheres to the theme. If you'd like to present, but do not have original research on the topic, please select several papers from either of the lists:

1. List 1

2. List 2

3. Three papers on SOS

* Better Agnostic Clustering Via Relaxed Tensor Norms. Pravesh K. Kothari and Jacob Steinhardt. STOC 2018

* The power of sum-of-squares for detecting hidden structures. Pravesh K. Kothari, Samuel B. Hopkins , Aaron Potechin , Prasad Raghavendra , Tselil Schramm and David Steurer. FOCS 2017.

* Outlier-Robust Moment Estimation Via Sum-of-Squares. Pravesh K. Kothari, David Steurer. STOC 2018

and then email one of the organizers, or speak to them at the seminar.

This is a research and learning seminar. When presenting a paper, speakers should answer the following questions:

  • What are the authors trying to do?

  • How is it done today?

  • What are the limitations?

  • What are the future research directions?

Date Speaker Title
Feb 5 Yingjie (Tom) Fei Exponential error rates of SDP for block models: Beyond Grothendieck's inequality
Feb 12 Mateo Diaz Convex Recovery of A Structured Signal From Independent Random Linear Measurements
Feb 19 February Break
Feb 26 No Seminar
Mar 5 Abby Turner Computational Complexity Versus Statistical Performance On Sparse Recovery Problems
March 12 No Seminar
Mar 19 Vasileios Charisopoulos Entrywise Eigenvector Analysis of Random Matrices with Low Expected Rank
March 26 Lijun Ding Leave-one-out approach for Matrix Completion: Primal and Dual analysis
Apr 9 Calvin Wylie Solving (most) of a set of quadratic equalities: Composite optimization for robust phase retrieval
April 16 Yang YuanSpurious Local Minima are Common in Two-Layer ReLU Neural Networks
Apr 23
Apr 30
May 7