Past Events

Heterogeneous and large volume of Electronic Health Records (EHR) data are becoming available in many healthcare institutes, which include diagnosis, procedures, medications, lab results, clinical notes, medical images, genetic information and etc. Such EHR data from millions of patients serve as huge collective memory of doctors and patients over time. How to leverage that EHR data to help caregivers and patients to make better decisions in future?

Finite alphabet signaling refers to commonly used discrete-constellation modulations in practical communication systems, such as PAM, PSK or QAM. In this talk, we will target at how to increase data rate or throughput via linear precoding in wireless systems and networks such as multiple-input multiple-output (MIMO) systems, multiple access channels, broadcast channels, wiretap channels, and cognitive radio networks.

This presentation focuses on a classical detection problem of binary signals corrupted by impulse noise modeled by a Middleton Class-A (MCA) distribution. This distribution is one of the most accepted models for impulse noise superimposed to additive white Gaussian noise. The optimum detector in such noise consists of optimum operations followed by a conventional combiner. Since the MCA model is expressed as a weighted linear combination of an infinite number of Gaussian densities, there is no closed-form solution for the optimum preprocessor.

The large-scale gathering and storage of personal data is raising new questions about the regulation of privacy.  On the technology side, there has been a flurry of recent work on new models for privacy risk and protection.  One such model is differential privacy, which quantifies the risk to an individual's data being included in a database.  Differentially private algorithms introduce noise into their computations to limit this risk, allowing the output to be released publicly.

Capacity projections from the International Mobile Telecommunications (IMT) have become outdated even before the next generation, i.e., the 4th generation (4G), wireless communications systems have been widely deployed. This is due to the fairly recent explosion and uptake of smart phones and smart services anytime/everywhere, and the thus associated data requirements.The trend is clearly towards splitting indoors and outdoors network designs, with focus on interoperation, support of mobility, high traffic levels and emerging bandwidth-intensive applications. Focusing e.g.

It is well-known that belief-propagation (BP) decoding of low-density parity-check (LDPC) codes is suboptimal and that the noise threshold of maximum-a-posteriori (MAP) decoding can be larger than the BP threshold.  Recently, Kudekar et al. proved that regular LDPC ensembles can be spatially coupled (SC) so that the BP noise threshold saturates to the MAP noise threshold of the original ensemble.  These SC ensembles are instances of LDPC convolutional (LDPCC) codes and the new proof explains an earlier observation by Lentmaier et al.

In the current work we deal with the problem of base station cooperation in the downlink of infinite wireless cellular networks. The positions of base stations are modeled by a Poisson point process. Each base station can choose to cooperate or not with exactly one of its Delaunay neighbours in order to provide service to a user located within its cell.  The cooperation protocol uses a variation of the so-called Willems' encoder and a fixed total transmission power per user is considered.

While information theory was initially developed specifically for the study of reliable communication systems, it has found broader applications in areas as diverse as biology, cryptography, or machine learning. More recently, motivated in part by the emergence of distributed cyber-physical systems, information theory has also been used to develop the basis of a theory of coordination over networks. The central idea behind the approach is to view the transmission of data between nodes in a network as a means to coordinate their behaviors, and not as an end in itself.

We study the trade-off between delivery delay and energy consumption in a delay tolerant network in which a message (or a file) has to be delivered to each of several destinations by epidemic relaying. In addition to the destinations, there are several other nodes in the network that can assist in relaying the message. We first assume that, at every instant, all the nodes know the number of relays carrying the packet and the number of destinations that have received the packet.

Increasingly, optimization problems in machine learning, especially those arising from high-dimensional statistical estimation tasks, involve a large number of variables. As regards the statistical estimation tasks themselves, methods developed over the past decade have been shown to have *statistical or sample complexity* that depends only weakly on the number of parameters, when there is some structure to the problem, such as sparsity. A central question is whether similar advances can be made in their *computational complexity* as well.