Past Events

The brain is simultaneously a biological object and a computing machine that represents and stores information about the external world, computes with this information, and generates adaptive behaviors in the world. Modern neuroscience is bringing sophisticated tools to observe neural codes. Our actual understanding of these codes is at a nascent stage, and opportunities to understand how the brain codes information and why it does so the way is does, abound. I'll talk about some fascinating neural codes and our work to theoretically analyze these codes.

Title: Towards a Mobile Content Marketplace: Proactive Caching and Pricing Strategies To view presentation slides from the day's event, view the PDF file.

Many emerging applications of machine learning involve time series and spatio-temporal data. In this talk, I will discuss a collection of machine learning approaches to effectively analyze and model large-scale time series and spatio-temporal data, including temporal causal models, sparse extreme-value models, and fast tensor-based forecasting models. Experiment results will be shown to demonstrate the effectiveness of our models in climate science and healthcare applications.

Millimeter wave wireless propagation and communications system design

In this talk, I will present two recent results in random matrix theory. In the first part of the talk, I will present a result that leads to analyzing the required number of labeled examples (also known as label complexity) of graph-based methods for semi-supervised learning. Graph-based methods have been quite successful in solving the semi-supervised learning problem, as they take into account the underlying geometry of the data.

Rate splitting and iterative decoding approaches for fading channels with imperfect CSI

In this work, Prof. Vangelis Markakis studies two standard multi-unit auction formats for allocating multiple units of a single good to multi-demand bidders. The first one is the Discriminatory Price Auction, which charges every winner his winning bids. The second is the Uniform Price Auction, which determines a uniform price to be paid per unit. Variants of both formats find applications ranging from the allocation of bonds to investors, to online sales over the internet, facilitated by popular online brokers.For these formats, Dr.

In this talk, we propose new coding techniques based on coset codes for communication over three multi-terminal channels including the three user discrete broadcast (3-BC) and interference channels (3-IC). Characterizing the performance of the proposed coding technique in an information theoretic framework enables us present new achievable rate regions. These new achievable rate regions strictly enlarge upon the current known largest that were derived over three decades ago.

IEEE GLOBECOM is one of two flagship conferences of the IEEE Communications Society (ComSoc), together with the IEEE ICC. Each year the conference hosts over 1,000 peer-reviewed technical papers and a cutting-edge industry program. The conference meets in North America and attracts roughly 2,000 leading scientists, researchers and industry practitioners from around the world. This year, Dr. Robert Heath and Dr.

In the matrix completion problem, one sees a few entries of an (approximately) low-rank matrix and hopes to (approximately) recover the entire matrix.  This problem has gotten a lot of attention recently, partly due to its applicability to recommender systems.  It is known that, by solving a convex program, the original matrix can be recovered with a number of observations that is linear in n.  However, for current analyses of faster algorithms (with runtime linear in n), the number of samples additionally depends at least quadratically on the *condition number* of the matrix.&n