Extreme Value Analysis (EVA) 2023

13th conference on Extreme Value Analysis, Probabilistic and Statistical Models and their Applications

The 13th International Conference on Extreme Value Analysis has been hosted by Bocconi University from June 26 to June 30 2023 in Milan.

Chairs of the local organising committee:

  • Simone Padoan (lead organizer);
  • Stefano Rizzelli;
  • Andrea Toreti.

The EVA conference aim is to bring together researchers in Extreme Value Theory, Methods and its Applications. The conference has concerned with a broad spectrum of topics ranging from 

  • Heavy-tailed phenomena; 
  • Time series; Large deviations;
  • Multivariate modelling; Rare events; 
  • Spatio-temporal processes;


  • Machine Learning; 
  • Graphical Modelling;
  • Big data;
  • Networks;

and has included many important applications such as

  • Energy Systems;
  • Climate;
  • Public Health;
  • Epidemiology;
  • Life Sciences;
  • Insurance and Finance.

The content of the website is inspired by the websites of the previous conference editions, to whose organizers we are grateful.

The local organizing committee comprises:

  1. Amir Khorrami, Università di Torino;
  2. Simone Padoan (chair), Bocconi University;
  3. Stefano Rizzelli (website manager), Università Cattolica;
  4. Andrea Toreti, European Commission, Joint Research Center.

The scientific committee comprises:

  1. Axel Bucher, Dusseldorf University;
  2. Valérie Chavez, Université de Lausanne;
  3. Richard Davis, Columbia University;
  4. Anthony Davison, EPFL;
  5. Clément Dombry (chair), Université de Franche-Comté;
  6. Ana Ferreira, University of Lisbon;
  7. Armelle Guillou, Université de Strasbourg;
  8. Enkelejd Hashorva, Université de Lausanne;
  9. Anja Janßen, Magdeburg University
  10. Rafal Kulik, University of Ottawa;
  11. Thomas Mikosch, University of Copenhagen;
  12. Natalia Nolde, The University of British Columbia
  13. Simone Padoan, Bocconi University;
  14. Gennady Samorodnitsky, Cornell University;
  15. Johan Segers, UCLouvain;
  16. Kirstin Strokorb, Cardiff University;
  17. Gilles Stupfler, Université d'Angers;
  18. Chen Zhou, Erasmus University.



EVA 2023 has been organised by Bocconi University. 

The conference brings together researchers in Extreme Value Theory, Methods and its Applications.

Its inception took place almost 40 years ago in Vimeiro, Portugal (1983). The photo below has been taken from the previous EVA conference edition:

First edition of the conference (courtesy of Feridun Turkman and Laurens de Haan)


Previous EVA conferences took place in:

EVA2023 program has included four plenary sessions:


  • Climate risk assessment - scope and guiding principles 


  • Fingerprint heatwaves and cold spells using large deviation theory



  • In memory of James Pickands
    •  Laurens de Haan (Erasmus University Rotterdam)
    • Guus Balkema (Amsterdam University)


EVA2023 program has inculded 20 invited sessions, concerning probability, statistics and applications.




Extremes of stochastic processes and random fields

Organizer: Krzys Debicki


  • Tomasz Rolski
  • Lanpeng Ji
  • Nikolai Kriukov

Long memory and extremes

Organizer: Yizao Wang


  • Shuyang Bai 
  • Bojan Basrak
  • Rajat Subhra Hazra

Extremes and stochastic geometry

Organizer: Christian Hirsch


  • Gilles Bonnet
  • Nicolas Chenavier
  • Takashi Owada

Extremes in networks

Organizer: Phyllis Wan


  • Tiandong Wang 
  • Claudia Kluppelberg
  • Fiona Sloothaak




Non-stationary extremes

Organizer: Axel Bücher


  • Ana Ferreira
  • John Einmahl
  • Jonathan Tawn

Spatio-temporal extremes

Organizer: Thomas Opitz


  • Carlo Gaetan
  • Raphael Huser
  • Manuela Brunner

High dimensional extremes

Organizer: Chen Zhou


  • Yi He
  • Stanislav Volgushev
  • Anne Sabourin

Multivariate extremes

Organizer: Anna Kiriliouk


  • Marco Oesting 
  • Michaël Lalancette
  • Jenny Wadsworth

Imperfect extremes

Organizer: Armelle Guillou


  • Gennady Samorodnitsky
  • Mikael Escobar-Bach
  • Yuri Goegebeur

Extremes and machine learning

Organizer: Olivier Wintenberger


  • Nicola Gnecco
  • Olivier Pasche
  • Jordan Richards

Time series and extremes

Organizer: Anja Janssen


  • Axel Bücher
  • Philippe Soulier
  • Marie-Christine Düker

Bayesian methods for extreme value analysis

Organizer: Simone Padoan


  • Ben Shaby
  • Stefano Rizzelli
  • Miguel de Carvhallo

Graphical models

Organizer: Claudia Kluppelberg


  • Phyllis Wan
  • Johan Segers
  • Carlos Amendola

Conditional relationships and sparsity for extremes

Organizer: Dan Cooley


  • Sebastian Engelke
  • Jeongjin Lee
  • Yan Gong




Finance, insurance, risk management

Organizer: Natalia Nolde


  • Debbie Dupuis
  • Armelle Guillou
  • Chen Zhou

Extremes for environment and energy

Organizer: Raphael Huser


  • Almut Veraart
  • Thordis Thorarinsdottir
  • David Bolin

Impact of extremes on human activities

Organizer: Maud Thomas


  • Olivier Lopez
  • Juliette Legrand
  • Jonathan Koh

Concurrent and recurrent climate extremes I

Organizer: Andrea Toreti


  • Ottmar Cronie
  • Erich Fisher
  • Carmen Alvarez-Castro

Concurrent and recurrent climate extremes II

Organizer: Andrea Toreti


  • Reik Donner
  • Arthur Hrast Essenfelder
  • Jakob Zscheischler

Climate related extremes and risk

Organizer: Jaroslav Mysiak


  • Riccardo Henin
  • Leone Cavicchia
  • Eirini Sartsali

Two tutorials of about 3 hours have been held (in parallel) in the afternoon of Friday 30 June 2023.

Details on the tutorials are provided here below.


Handling asymptotic theories in extremes: (multivariate) tail empirical and tail quantile processes


Instructor: Chen Zhou (Erasmus University Rotterdam)

Description: Extreme value statistics often rely on selected observations in the tail of a (multivariate) distribution. Observations are selected by using either the Peaks-over-threshold (POT) or the block maxima (BM) approaches. Statistical estimators are regarded as functional transformation of empirical or quantile processes formed based on such tail observations, the tail empirical and tail quantile processes. To establish asymptotic theories for most estimators, the fundamental tool is their asymptotic behavior.

In this tutorial, we start from asymptotic behavior of the weighted tail empirical and tail quantile processes based on independent and identically distributed observations. Then we discuss how to use them in proving asymptotic theories of typical extreme value estimators. The tutorial will also cover advanced tools such as multivariate tail empirical process and tail empirical and tail quantile processes based on non-i.i.d. data, e.g. time series and non-stationarity.

At the end of the tutorial, one should be able to use these tools to establish asymptotic theories of new extreme value type estimators developed in their own work.


Tutorial on Statistical Computing on Extremes with R


Instructors: Léo Belzile (HEC Montréal) and Thomas Opitz (INRAE Avignon)

Description: This workshop will review R implementations of a variety of techniques in statistical analysis. The focus of the first part of the workshop will be on univariate extremes, including likelihood-based, Bayesian and nonparametric methods for both peaks-over-threshold and block maxima approaches, with a foray into nonstationary extremes. The second part of the workshop will concentrate on conditional extremes and time series. These two sessions will be supported by a recent review paper co-authored by Léo and Thomas.


The Springer journal Extremes has organized a competition for Best Student Paper Award at EVA 2023 in Milan.

The corresponding author of the best paper will be awarded a prize of 500 € sponsored by Springer-Verlag with the option of publishing the paper in Extremes (after a review).


Manuel Hentschel (Université de Genève)

Statistical Inference for Hüsler-Reiss Graphical Models Through Matrix Completions

(Joint work with Sebastian Engelke and Johan Segers)

The Scientific committee has organized a competition for Best Poster Award at EVA 2023 in Milan. 


Winners (in alphabetic order)



Paolo Victor Redondo (King Abdullah University of Science and Technology)

for presenting the poster 

Spectral Transfer Entropy and its Application to Extremal Brain Causal Connectivity Networks

(joint work with Raphaël Huser and Hernando Ombao)


Torben Staud (Heinrich-Heine-University Düsseldorf)

for presenting the poster

U-statistics of extremes based on disjoint and sliding block maxima

(joint work with Axel Bücher)


Victor Verma (University of Michigan)

for presenting the poster

Forecasting Solar Flares with Extreme Value Theory

(joint work with Yang Chen and Stilian Stoev)

EVA 2023 announces a data challenge competition, organised by  

Detailed instructions (along with some further clarifications) and data can be downloaded here.


Team Yalla (King Abdullah University for Science and Technology)


In this area, it is possible to download the conference program

For any inquiry, write to eva23.board@unibocconi.it.