This talk reviews some recent work on (unpenalized) linear re- gression M-estimators in high-dimensions. Extending the seminal work of Peter Huber, Steve Portnoy and others to the setting where n, the number of ob- servations, is large and comparable to p, the number of predictors, we obtain updated results for …
Proteins can be thought of as the machines inside cells, participating in nearly all biological processes. By physically binding to and interacting with one another, proteins form pathways dedicated to specific functions and vast networks. Mapping these pathways is critical for understanding how their disruption causes disease, but biological experiments …
In many scientific disciplines, new technologies are enabling researchers to obtain measurements of unprecedented scale and resolution. These huge data sets present many new challenges and opportunities for the experimental scientists generating the data and for researchers like those in the SILO community developing data analyses. In this talk, three …
Consider a generic r-dimensional subspace, and suppose that we are only given projections of this subspace onto small subsets of the canonical coordinates. In this talk I will show the necessary and sufficient conditions on such subsets to guarantee that there is only one r-dimensional subspace consistent with all the …
With the recent data revolution, statisticians are considering larger datasets, more sophisticated models. As a consequence, it is important to seek methods that are computationally efficient as well as statistically powerful. We will show how hypotheses from theoretical computer science on the hardness of some problems in average case can …
The advent of micro and nanotechnology for interfacing with the nervous system has opened up a wide range of experimental platforms for getting information into and out of the living brain. This talk will provide an overview of the different approaches for recording and stimulating large numbers of neurons in …
Topological data analysis looks at data from a rather unique angle. Good news: it may provide additional information to traditional machine learning, hence benefiting downstream applications. Bad news: Let K be the social network over topologists and machine learners, then betti0(K)=2. Why this means they didn’t talk to each other …
This talk gives an introduction to a recently established link between the geometry of numbers and mixed integer linear optimization. The main focus is to provide a review of families of lattice-free polyhedra and their use in the context of deriving and explaining cutting planes for mixed integer programs. This …
We provide a high level overview of the information theoretic study of two-way networks. We start by introducing the capacity / capacity region of a channel / network, respectively. While we know the capacity of one-way point-to-point channels where there is one source and one destination, we do not know …
The problem of low rank estimation naturally arises in a number of functional or relational data analysis settings, for example when dealing with spatio-temporal data or link prediction with attributes. We consider a unified framework for these problems and devise a novel penalty function to exploit the low rank structure …