Department of Mathematical Sciences Landscape seminar
Invited speakers give an overview of a topic of general mathematical interest, aimed at staff and postgraduate students. Others are also welcome to attend.
Time and location
The Landscape seminar normally runs on Fridays from 15:15 to 16:05 in the Wolfson Lecture theatre (4W 1.7) with refreshments served afterwards. There is roughly one Landscape seminar per month in semester time.
7 February 2020 - Joachim Krug, University of Cologne
(invited by the Probability group)
Title: The paths not taken: Evolutionary accessibility in random fitness landscapes
Abstract: Biological evolution can be conceptualized as a search process in the space of gene sequences guided by the fitness landscape, a mapping that assigns a measure of reproductive value to each genotype. For a long time after their introduction almost a century ago, fitness landscapes have played a largely metaphorical role, but in recent years the increasing availability of empirical data sets has invigorated the field and motivated new mathematical questions. After a general introduction into the biological context, the talk will define probabilistic models of fitness landscapes and show interesting links with percolation theory.
6 March 2020 - Ruth Baker, University of Oxford
(invited by the Centre for Mathematical Biology) Title and abstract to appear
27 March 2020 - Jon Wilkening, University of California at Berkeley
(invited by the Applied and interdisciplinary mathematics group) Title and abstract to appear
24 April 2020 - Diane Maclagan, University of Warwick
(invited by the Algebra and geometry group) Title and abstract to appear
11 October 2019 at 12:15 - Martin Gander, University of Geneva
(invited by the Numerical Analysis group)
Title: Seven things I would have liked to know when starting to work on Domain Decomposition
13 December 2019 at 15:15 - Stijn Vansteelandt, Ghent University
(invited by the Statistics group)
Title: Causal machine learning: challenges, solutions and improvements
Abstract: The machine learning literature has offered enormous contributions on how to predict outcomes based on possibly high-dimensional predictors or features. Its role in evaluating the effects of treatments, exposures, interventions, or policies on outcome has been more minor so far. This is surprising if one considers that many, if not most, empirical studies are primarily aimed at inferring cause-effect relationships. In this talk, I will therefore focus on the use of machine learning for the evaluation of (causal) treatment effects. This turns out to be a challenging task: while the prediction performance of a given machine learning algorithm can be measured by contrasting observed and predicted outcomes, such evaluation becomes impossible when machine learning is used for treatment effect estimation since the true treatment effect is always unknown. Mathematics (in particular, asymptotic mathematical statistics) therefore unavoidably plays an essential role in the construction of causal machine learning techniques. In this talk, I will develop both intuitive and formal insight into this. I will demonstrate why naive use of existing machine learning algorithms is problematic, will give a gentle introduction to pioneering work on Targeted Learning and on Double Machine Learning, and will discuss recent improvements of these techniques. Throughout the talk "machine learning" will be considered in the broad sense as any algorithm that uses data to "learn" a proper model for the data, thus including (though not being limited to) routine variable selection procedures. The talk will be accessible to attendees without a detailed understanding of machine learning algorithms.