LIDS & Stats Tea

Tea talks are 20 minute long informal chalk-talks for the purpose of sharing ideas and making others aware about some of the topics that may be of interest to the LIDS and Stats audience. If you are interested in presenting in the upcoming seminars, please email lids_stats_tea[at]mit[dot]edu

September 15, 2021

Sensitivity and Counterfactual Analysis of Structural Econometric Models

Yaroslav Mukhin (LIDS)

I propose a framework to characterize the sensitivity of estimates and counterfactuals to the distributional assumptions about latent variables in structural econometric models. I characterize the lower and upper bounds on the counterfactual as the...

September 22, 2021

In Defense of Fluid Democracy

Manon Revel (LIDS & IDSS)

In response to the frustration with the current democratic institutions that failed to adapt to the evolving society, alternative decision-making processes emerged in recent years. Requests for more representativity of the citizenry resulted in the...

September 29, 2021

Predicting Failure Cascades in Large Scale Power Systems via the Influence Model Framework

Xinyu Wu (LIDS)

Large blackouts in power grids are often the consequence of uncontrolled failure cascades. The ability to predict the failure cascade process in an efficient and accurate manner is important for power system contingency analysis. In this work, we...

October 6, 2021

Online Learning in Unknown Markov Games

Yi Tian (LIDS)

Sequential decision-making in a multi-agent environment arises frequently in the real world. We consider online learning in unknown Markov games: 1) a Markov game is used to model the multi-player dynamic environment, 2) online learning means that...

October 13, 2021

Causal Matrix Completion

Anish Agarwal (LIDS)

Matrix completion is the study of recovering an underlying matrix from a sparse subset of noisy observations. Traditionally, it is assumed that the entries of the matrix are “missing completely at random” (MCAR), i.e., each entry is revealed at...

October 20, 2021

Proximal-Primal-Dual Algorithms for Constrained Nonconvex Optimization Problems

Jiawei Zhang (LIDS)

In this talk, we introduce our recent works about proximal-primal-dual algorithms for constrained nonconvex optimization. The augmented Lagrangian method (ALM) and the alternating direction method of multipliers (ADMM) are popular for solving...

October 27, 2021

Differential Geometry Methods in Electric Energy Systems with Distributed Renewable Energy Resources

Dan Wu (LIDS)

In recent years extreme weather conditions such as serious floods in China and Europe, long-lasting wild fires in America and Australia, extreme winter cyclones in the southern part of the U.S., and extreme heat in the Arctic circle occurred more...

November 3, 2021

Habits in Consumer Purchases: Evidence from Store Closures

Amir Tohidi Kalorazi (LIDS & IDSS)

Our goal is to measure the effect of habit formation on consumers’ behavior, in particular, the widespread experience of shopping in a store. In-store shopping is a repetitive behavior happening in the same context, so it has a great potential of...

November 10, 2021

Determining Influential Edges with Higher Order Information

Arnab Kumar Sarker (IDSS)

Social capital, which refers to the resources an individual has due to their position in a social structure, is critical for individual and organizational success in many settings.  In this work, we focus on the task of identifying important edges...

November 17, 2021

Learning Partial Differential Equations in Reproducing Kernel Hilbert Spaces

George Stepaniants (Mathematics)

The rapid development of data-driven scientific discovery holds the promise of new and faster methods to analyze, understand, and predict various complex phenomena whose physical laws are still beyond our grasp. Of central interest to this...


December 1, 2021

Learning to Delegate for Large-scale Vehicle Routing

Zhongxia (Zee) Yan (LIDS)

Vehicle routing problems (VRPs) have enjoyed ample applications in logistics and ride-hailing services around the world for decades. While determining the optimal solution to a VRP is NP-hard, powerful heuristic solvers such as LKH-3 and HGS find...

December 8, 2021

Low-Rank Sum of Squares Has No Spurious Local Minima

Chenyang Yuan (LIDS)

We study the problem of decomposing a polynomial into a sum of r squares by optimizing the objective fₚ(u) = ‖∑ᵢ uᵢ² - p‖². This objective is nonconvex and is equivalent to the rank-r Burer-Monteiro factorization of a semidefinite program (SDP)...


December 15, 2021

Introducing Discrepancy Values of Matrices and Their Applications

Pourya Habib Zadeh (LIDS)

We introduce discrepancy values, quantities that are inspired by the notion of spectral spread of Hermitian matrices. In particular, the discrepancy values capture the difference between two consecutive (Ky-Fan-like) pseudo-norms that we also...