Event Details

This workshop brings together Emeritus Distinguished Professor Jon Tawn's past PhD students, collaborators and academic friends.

Tuesday 30th June schedule:

• 13.30-14.00 Coffee/Arrival (LUMS Hub)
• 14.00-14.10 Opening (LUMS Lecture Theatre 4)
• 14.10-15.00 Stuart Coles: Is Jonathan Tawn an under-achiever?

Sure, 50-plus PhD students in an academic career is a bit of a milestone. But did Jonathan achieve as much as he could, or indeed should, have? More precisely, if he hadn't opted for early retirement, what could he realistically have accomplished? We'll use some of the most sophisticated extreme value tools and software available to examine this question. The answer, at least by one version of the analysis, puts his apparent "success" in rather poor light and leaves us wondering what might have happened if he'd only stuck at it a bit longer.

• 15.00-15.15 Jess Spearing: Life as a Jon Tawn PhD Student… and beyond.
• 15.15-15.30 Hugo Winter: Everybody’s Got Something To Hide Except Me and My Tawny

In this unserious talk about my time working with Jon from 2011 onwards I aim to mix two of his lifelong passions: extreme value theory and The Beatles! Some consider Lennon and McCartney to be the premier songwriting duo of their time (if not ever), but we shall see that for a brief period from 2011-2015 they were threatened by a new partnership in Winter and Tawn. With a string of moderate hits and a fantastic mop-top look (well at least one of them), what could have become of this pair had the changing times not curtailed their relationship and led to the inevitable solo careers?

• 15.30-15.50 Break
• 15.50-16.05 Eleanor D’Arcy: From Modelling Extremes to Measuring Them: Statistical Design of a National River Flow Network.

Jon was my PhD supervisor from 2020 - 2024. During this time, we developed research that built on work from his own PhD, focusing on the estimation of extreme sea levels using a joint probabilities methodology. Jon’s supervision extended far beyond technical guidance: he strongly encouraged me to engage in research competitions, outreach activities and collaborations with organisations outside academia. One such organisation was the Environment Agency, where I now work as a Senior Statistical Advisor on the Natural Capital and Ecosystem Assessment programme.

I will begin this talk with an overview of my time as Jon’s PhD student and how his support shaped my development as a statistician. In the later stages of my PhD, we derived extremal properties for a class of models suitable for extreme river flows. In my current role at the Environment Agency, I am the statistical lead for the design of a new national river flow monitoring network, marking a shift from modelling extreme river flows to designing how they are monitored. I will introduce this network and its statistical design, with the aim of highlighting a rich new dataset that may be of direct relevance to researchers working on environmental extremes.

• 16.05-16.15 Marian Scott and Daniela Castro-Camilo
• 16.15-17.05 Thomas Opitz: The long and winding road of dependence modelling for extremes.

Jon's work has been foundational and highly influential for descriptive and model-based analysis of dependence of extreme events. The aim of this talk is to give a short overview with focus on Jon's contributions, spanning from more traditional models based on multivariate regular variation to more recent approaches allowing for more flexibility in characterizing joint tail decay rates. A particular focus will be on the conditional extremes framework, with some examples of its use in ongoing projects on environmental extremes highlighting the versatility of this approach.

• 18.00 Evening meal - Grizedale Bar - please note that the bar will not be open, although drinks will be available with food. For those who wish to socialise after, we would recommend either the Lancaster House Hotel or going into the city centre.

Wednesday 1st July schedule (LUMS Lecture Theatre 4):

• 09.30-09.45 Conor Murphy: Threshold selection, induced seismicity & lots of scribbles!

My PhD research, carried out under the supervision of Jon Tawn, focused on developing methodology for a fundamental problem in extreme value analysis: threshold selection in the peaks-over-threshold framework.

In this talk, I will provide an overview of the methodology developed for automated threshold selection in the iid setting, along with approaches to propagate threshold uncertainty into inference. I will then describe how these ideas were extended to the more complex setting of induced seismicity, where spatio-temporal models are required to account for evolving measurement networks and incomplete observations.

Along the way, I will reflect on the development of these projects during my PhD and some highlights of working with Jon, whose ability to reduce complex problems to a few simple scribbles on a page was key to the success of our work.

• 09.45-10.05 Anthony Ledford: What's the score with AI?

At the risk of disappointing some, this talk won't be about football. Or artificial intelligence. I'll instead cover how Jon steered me on an unexpected journey into asymptotic independence that started 34 years ago.

• 10.05-10.55 Dan Cooley: Modeling Extremes in High Dimension, the TPDM, and Vector Spaces.

This talk will explore how much can be learned about high-dimensional extremes when viewed through pairwise relationships. Since fully characterizing the extremal dependence structure in high dimensions is difficult, summarizing extremal dependence via pairwise measures, whose estimation is straightforward, can provide actionable dependence information for modeling. This talk will demonstrate how the tail pairwise dependence matrix can be used to explore and model extremes in high dimensions.

The talk will then turn attention toward connecting standard ideas about vector spaces and inner products to regular variation and the TPDM. One challenge is that collections of regularly varying random variables cannot be assumed to be jointly regularly varying. In ongoing work, we define a vector space of regularly varying random variables, and assign a norm that is related to the `scale' of the random variable's tail. Current work is to show that the TPDM corresponds to an inner product in the case when the tail index alpha = 2.

• 10.55-11.05 Group Photo (Bonington Steps)
• 11.05-11.15 Break (Postgraduate Statistics Centre A floor)
• 11.15-11.30 Emma Simpson: Some highlights of working with Jon: A magical mystery tour.

I had the privilege of being supervised by Jon during my PhD (2015-2019) and postdoc (2019-2021), as well as collaborating on projects after leaving Lancaster. From various projects on multivariate extremes to co-organising the EVA data challenge in 2023, I will discuss some of my memories and things I’ve learned from working with Jon over the years. I’ll also highlight some of the influences I see Jon’s work having on my research going forward.

• 11.30-11.50 Mark Dixon: Patient capital: Sparx, Oxygen House and Atass Sports 25 years in.
• 11.50-12.40 Anthony Davison
• 12.40-12.50 Close
• 12.50-14.00 Lunch (Grizedale Bar)