Usual time: Thursdays 14:00-15:00
Location: Room 102, Department of Statistical Science, 1-19 Torrington Place (1st floor). Some seminars are held at different locations and at different times. Please click on the abstract for further details.
- 16 January 2020: Dr. Wei Qian (University of Cambridge)
Generalized disconnection exponents for Brownian loop-soups
We study the question of whether there exist double points on the boundaries of clusters in Brownian loop-soups — an object introduced by Lawler and Werner in 2004. This question is closely related to our earlier works (with Werner) on the decomposition of Brownian loop-soup clusters. More concretely, we introduce a notion of disconnection exponents which generalizes the Brownian disconnection exponents derived by Lawler, Schramm and Werner in 2001. By computing the generalized disconnection exponents, we can predict the dimension of multiple points on the cluster boundaries in loop-soups. However, for the critical intensity of loop-soup, the dimension of double points on the cluster boundaries appears to be zero, leaving the open problem of whether such points exist for the critical loop-soup.
- 06 February 2020: Dr. Anne Presanis (University of Cambridge)
Value of information in a Bayesian evidence synthesis to estimate HIV prevalenceAnnual estimation of the number of people living with HIV in England, particularly those who are unaware of their infection, has, for several years, been based on a Bayesian model that combines evidence from multiple sources of data. For several demographic and risk groups, the the number of people in each group, the prevalence of HIV, and the proportion of HIV infections that are diagnosed are all estimated. This is an example of a “multiparameter evidence synthesis”, where the quantities of interest cannot be estimated directly, but can be inferred indirectly through a network of model assumptions. Model assessment is important for any model, but particularly for the complex probabilistic models entailed by evidence synthesis: it is important to know which model parameters most affect the estimate, in particular which ones drive the uncertainty in the estimates, in order to prioritise what further data should be collected to reduce the uncertainty. These questions can be addressed by a Value of Information (VoI) analysis, in which we estimate expected reductions in uncertainty, expressed as a loss, from learning specific parameters or collecting data of a given design. VoI methods are illustrated for the HIV example, and a consequent development to the HIV model to make more comprehensive and effective use of the available data is described, resulting in estimates of recent trends in HIV prevalence in England. The proportion of people living with HIV aware of their infection steadily increased from 84% (95% credible interval 77-88%) to 92% (89-94%) over 2012-2017, corresponding to a halving in the number of undiagnosed infections from 13,500 (9,800-20,200) to 6,900 (4,900-10,700), and reaching the UNAIDS 90-90-90 targets in 2016.
- 13 February 2020: Dr. Benjamin Dadoun (University of Bath)
Self-similar growth fragmentations as scaling limits of Markov branching processes
We provide explicit conditions, in terms of the transition kernel of its driving particle, for a Markov branching process to admit a scaling limit toward a self-similar growth fragmentation with negative index. We also derive a scaling limit for the genealogical embedding considered as a compact real tree.
- 27 February 2020: Dr. Mine Cetinkaya-Rundel (University of Edinburgh)
The art and science of teaching data science
Modern statistics is fundamentally a computational discipline, but too often this fact is not reflected in our statistics curricula. With the rise of data science it has become increasingly clear that students want, expect, and need explicit training in this area of the discipline. Additionally, recent curricular guidelines clearly state that working with data requires extensive computing skills and that statistics students should be fluent in accessing, manipulating, analyzing, and modeling with professional statistical analysis software. In this talk we introduce the design philosophy behind an introductory data science course, discuss in progress and future research on student learning as well as new directions in assessment and tooling as we scale up the course.
- 05 March 2020: Dr. Patrick Rubin-Delanchy (University of Bristol)
- 12 March 2020: Dr. Eva Cantoni (University of Geneva)
- 26 March 2020: Dr. Tra My Pham (MRC Clinical Trials Unit at UCL)