LIDS Seminar Series

September 22, 2015

Optimal Resource Allocation to Control Epidemic Outbreaks in Networked Populations

Victor Preciado (University of Pennsylvania)

We study the problem of controlling epidemic outbreaks in networked populations by distributing protection resources throughout the nodes of the network. We assume that two types of protection resources are available: (i) Preventive resources able...

September 29, 2015

Some Limitations and Possibilities Toward Data-Driven Optimization

Yaron Singer (Harvard University)

As we grow highly dependent on data for making predictions, we translate these predictions into models that help us make informed decisions.  But how do the guarantees we have on predictions translate to guarantees on decisions? In many cases, we...

October 6, 2015

Applying Machine Learning in Online Revenue Management

David Simchi-Levi (MIT)

In a dynamic pricing problem where the demand function is unknown a priori, price experimentation can be used for demand learning. In practice, however, online sellers are faced with a few business constraints, including the inability to conduct...

October 20, 2015

Teaching an Old Code a New Trick

Rudi Urbanke (EPFL)

Our digital lives depend heavily on our ability to efficiently and reliably transmit information over long distances. It is therefore not surprising that much effort has been dedicated to devising clever schemes to accomplish this. I will go back in...

October 27, 2015

Dynamics of Swarms and Opinions via Discrete Positive Systems

Ulrich Krause (Universität Bremen)

In the talk a general model for swarm formation of birds (or other agents) will be presented. Swarm formation means that birds approach asymptotically the same velocity, whereby distances among them do converge. The main result offers conditions on...

November 3, 2015

How to Learn Probability Without Learning

Young-Han Kim (Univ. of California, San Diego)

As Laplace famously asked ``What is the probability that the sun will rise tomorrow?,'' inferring the probability underlying a given data sample is at the core of statistics. In this talk, I will present a general principle of assigning...

November 17, 2015

Hiding the Source of a Rumor in Anonymous Messaging

Sewoong Oh (University of Illinois)

Anonymous messaging platforms, such as Secret, Whisper and Yik Yak, have emerged as important social media for sharing one's thoughts without the fear of being judged by friends, family, or the public. Further, such anonymous platforms are crucial...

November 18, 2015

Train Faster, Generalize Better: Stability of Stochastic Gradient Descent

Ben Recht (Univ. of California, Berkeley)

The most widely used optimization method in machine learning practice is the Stochastic Gradient Method (SGM).  This method has been used since the fifties to build statistical estimators, iteratively improving models by correcting errors observed...

November 24, 2015

Coupling, Entropy and Costa's Corner-Point Conjecture

Yury Polyanskiy (MIT)

In this talk I will describe a new method for bounding mutual information. It consists of two steps. First, we show that under regularity conditions two high-dimensional random variables that are close in terms of expected distance necessarily have...

December 1, 2015

Learning from Geometry

Robert Calderbank (Duke University)

Deep neural networks have proved very successful in domains where large training sets are available, since their capacity can be increased by adding layers or by increasing the number of units in a layer. When the number of training samples is small...