Abstract

Research: Nice detectors for ugly lattices, Francisco Monteiro

This talk shows that it is useful to approximate a given random lattice by one with a structure that allows a trellis representation and thus trellis decoding. The problem of finding a quasi-orthogonal sublattice in a lattice is central for the solution to the problem. That problem is known in theoretical computer science as the Quasi-Orthogonal Subset problem (QOS) but had never been thoroughly researched as it was thought a theoretical question lacking applications to real problems. The solution of this problem was investigated as it is the key to design efficient trellis-based receivers for MIMO communication systems.

Research: Bringing The Next Billion Online: Computing Research For Developing Regions, Ripduman Sohan

This talk introduces the idea of technological research for developing regions in three parts. I begin by outlining the case for this research and highlight the (interesting and different) challenges associated with carrying out technical research in developing countries. I then provide some (varied) examples of such work highlighting its impact. I conclude by outlining some immediate areas of interest that overlap with the DTG ’s areas of expertise.