MSI Weekly Bulletin - Week starting Monday 21 August, 2006

Moduli spaces of Riemann surfaces with boundary introduced by C. C. Liu and Fukaya et al. give rise to explicit homotopy formulas in open/closed topological field theory (analogous to Massey products in rational homotopy theory). We give some examples, and show how these moduli spaces allow one to understand recent explicit calculations in the B-model.

The classification of 2-dimensional surfaces and the uniformisation theorem for the Riemannian geometries that they can carry provide a beautiful link between topology, Riemannian and conformal geometry as well as complex analysis. It has been a longstanding challenge to gain a similar understanding of all possible topological and geometrical shapes on 3-dimensional manifolds. Since the introduction of Ricciflow by Richard Hamilton it has become clear that parabolic geometric partial differential equations such as the Ricciflow for Riemannian metrics on abstract manifolds and the mean curvature flow for hypersurfaces in an ambient space can be used to deform given geometric objects into simpler, more easily recognised shapes. Hamilton proposed and initiated a method of how to use Ricciflow in combination with surgeries to break up a given 3-manifold into finitely many pieces, each of which is from a standard list conjectured by Thurston. Following breakthrough work of Perelman on the interlink of analytical and geometrical properties of Ricciflow recent detailed publications claim a full account of this program including a proof of the Poincare conjecture. The lecture describes Ricciflow and mean curvature flow with surgeries and tries to point out key strategies of the proof linking analysis and geometry in a beautiful way.

Techniques for matching, linking or integrating data are becoming increasingly popular in many organisations. While traditionally used mainly in health and statistics, today data linkage is increasingly being applied in and between government organisations to improve outcomes in taxation, census, immigration, social welfare, in crime and fraud detection, and in the assembly of terrorism intelligence. Many businesses routinely deduplicate and link their data when compiling mailing lists, and databases containing customer data are commonly sold for marketing purposes. Today, data linkage not only faces computational and operational challenges due to the increasing size of data collections and their complexity, but also privacy and confidentiality challenges due to growing concerns by the general public about their personal information being linked and shared between organisations. In this talk I will (1) present a short history of data linkage, (2) provide an overview of various innovative linkage techniques that have been developed in the last few years, and (3) discuss the core technical research areas that need to be addressed in order to make large scale data linkage both feasible as well as acceptable by the general public. Biography: ---------- Dr Peter Christen is a lecturer at the Department of Computer Science at the Australian National University. He received his Diploma in computer science engineering from the ETH Zurich (Switzerland) in 1995 and his PhD in computer science from the University of Basel (Switzerland) in 1999. His research interests are data mining (especially data preprocessing and data linkage), high-performance computing, and most recently security and privacy preservation (in the context of data linkage and health informatics). In the last four years his research has concentrated on the project "Investigation and Development of Parallel Large Scale Record Linkage Techniques", an ARC Linkage project conducted in collaboration with and partially funded by the NSW Department of Health.