Code Quest - Fluid Flow #1: Reading the Water

Welcome to our first Code Quest!

Are you a Deep Learning enthusiast interested on how to approach complex data challenges? If so, we invite you to join CodeQuest, a collaborative event where you'll have the chance to both learn a cool new scientific topic and expand your Machine Learning skills. This edition's challenge is entitled "Fluid Flow" and it's brought to you in collaboration with Inductiva Research Labs and Microsoft. Inductiva Research Labs has made publicly available a dataset on Fluid Movements and challenges you to predict the movement of those fluid particles. Since it is always hard to start, especially with such a complex ML task, we prepared four virtual workshops given by experienced ML practitioners to present the problem as well as the research done around its solution. This event consists of the first virtual workshop.

Summary:
In this first talk, we will introduce the Fluid Flow problem and the Fluid Cube dataset produced by Inductiva Research Lab based on Smoothed Particle Hydrodynamics. We will go through its structure and exploit some cool problems around it with the ultimate goal to use Graph Neural Networks in order to mimic smoothed particle hydrodynamics simulations. Finally, we will also discuss some modelling approaches and start outlining solutions for this problem

Speaker:
Maria finished her master's degree in physics engineering at the University of Coimbra in 2020 and she's been working in the field of data science and machine learning ever since. During this period she dedicated most of her time to applying machine learning for enhancing numerical methods in the field of hydrodynamics at Inductiva Research Labs. Currently, she's working at Colab +Atlantic where she applies machine learning methods to a weather forecasting and downscaling problem.