skip to content

 

 

Monday 24th October 2022

Isaac Newton Institute

Cambridge,
United Kingdom

Background

Dispersive Hydrodynamics is a mathematical framework originally developed to describe multiscale nonlinear wave phenomena in dispersive media. This framework has further expanded and currently covers a range of phenomena including both dynamic and stochastic aspects of wave propagation. It is now a vibrant and continuously developing field, with a variety of physical applications.
 
In weather and climate forecasting, physical models of the Earth’s atmosphere and oceans are incorporated with observational data to generate predictions. Rogue waves – also known as freak waves – represent an example of a particularly challenging subject of study in the modelling of nonlinear waves. Rogue waves appear as ephemeral sudden giant waves, observed for instance in the ocean, and seem to have no known origin and a potentially high destructive impact on life or human infrastructures. Dispersive Hydrodynamics offers a key for the theoretical understanding of wave dynamics underpinning the formation of extreme phenomena, as well as the potential to help uncover the mystery behind rogue waves.
 
From an interdisciplinary perspective, it is natural to inquire how investigations in Dispersive Hydrodynamics may interact with other fields and reflect on the mutual benefits of knowledge transfer from one field to the other. This initiative will focus particularly on how Dispersive Hydrodynamics can be embedded in a data driven context and identify mathematical tools which could help address open problems in data processing and machine learning.
 
Artificial intelligent (AI) devices and Machine Learning algorithms are indeed pervasive and ubiquitous in technological applications. Machine learning has been proven to be an effective tool for scientific investigations allowing to discover non intuitive patterns and correlations in large and unstructured data sets. However, it is generally acknowledged that due their complexity, AI algorithms provide answers to problems but often the underpinning mechanisms leading to such answers are yet to be discovered or fully understood.
 
The interaction between Dispersive Hydrodynamics and modern AI is expected to develop at two levels. On one hand, research in Dispersive Hydrodynamics involves the study of a variety of models for nonlinear wave phenomena, and machine learning algorithms could be effectively used as a tool for model validation through the use of extensive data sets. On the other hand, Dispersive Hydrodynamics provides a general and powerful mathematical framework to understand a variety of complex phenomena in nonlinear dynamical systems and collective phenomena. This framework allows one to describe macroscopic phenomena in complex random systems where emerging collective behaviours provide a conceptual basis for building artificial learning devices.
 
This knowledge exchange day is part of a six month research programme at the Isaac Newton Institute on Dispersive hydrodynamics: mathematics, simulation and experiments, with applications in nonlinear waves. In particular, this event will directly follow on from Workshop 4: Statistical mechanics, integrability and dispersive hydrodynamics, which focuses on different aspects of nonlinear dispersive hydrodynamics, randomness and statistical mechanics, and integrable turbulence.
 

Aims and Objectives

The aim of this event is to bring together mathematicians and scientists working at the forefront of Dispersive Hydrodynamics and its applications, with end users from industry to further investigate potential connections. In particular, this event will focus on applications of Dispersive Hydrodynamics related to randomness and extreme wave events in nonlinear dispersive waves (such as rogue waves in the ocean or in optical fibres) as well as the connections of Dispersive Hydrodynamics with random networks and machine learning.
 
Combining machine learning techniques and the most advanced mathematical modelling of nonlinear and complex systems for making effective and reliable predictions on the occurrence of extreme events and related subsequent scenarios will be of great benefit for the developing area of Dispersive Hydrodynamics.
 
This workshop will feature a series of talks that cover two main session themes:

  1. Weather & Climate Forecasting and Oceanography
  2. Random Networks and Machine Learning

 
A Provisional Programme will be made available in the coming weeks.
 
We anticipate the event will start with registration at 09.30 on Monday 24th October and talks with Q&A until 16:30, followed by a networking reception at 16:30-17:30. All timings are GMT.
 

Posters

There will be the opportunity to present a poster. If you would like to express an interest in presenting a poster, please indicate this in the registration form.
 

Registration and Venue

This event is free to attend and registration is now open. To register, please follow the link in the lefthand panel.
 
The workshop will take place at the Isaac Newton Institute for Mathematical Sciences in Cambridge, United Kingdom. Please visit the Isaac Newton Institute website for further information about the venue.