Starts: 30th July Sat 11am - 12noon ET
Gain an in-depth understanding of the theory behind two state-of-the-art models for weather forecasting: FourCastNet by NVIDIA and NowCasting by DeepMind and learn how to incorporate physics-based knowledge in ML forecasting models using NVIDIA Modulus platform.
This engagement requires a lot of motivation, a few hours per week time commitment, and some prior experience in ML and forecasting.
Mahdi Torabi | PhD | Lead
❗ Meeting link: https://meet.jit.si/timeseries/weatherforecasting
❗ Weekly meeting time: Satursdays, 11am - 12noon ET
❗ Coordination channel: dg-time-series. If issues joining, drop an email to ‘firstname.lastname@example.org’.
The topic of this working group will be Weather Forecasting and it will consist of the following steps:
- Learning the theory behind two cutting-edge ML models for weather forecasting
- Learning how to work with NVIDIA Omniverse and Modulus platforms that allow one to incorporate physical knowledge of for example weather in ML models
- Create a reusable knowledge base, in the form of detailed learning notes/recipes, that can be used by participants of future working groups
- [stretch] use these platforms to forecast weather at a new geographic location
1) Join the coordination slack channel
2) Please go over the following items, write one paragraph on how you would like to contribute to the WG, and send it via a slack message to the DG lead.
a) Skilful precipitation nowcasting using deep generative models of radar [https://www.nature.com/articles/s41586-021-03854-z]
b) FourCastNet: A Global Data-driven High-resolution Weather Model using Adaptive Fourier Neural Operators [https://arxiv.org/abs/2202.11214]
Session #1 - #3:
- Get to know everyone!
- We will introduce the problem & why it matters.
- End of session #3 → Apply to join the Core Team
- Energy & competency are both important, but energy is a little more.
- You can still keep up with the WG if you don’t make the Core Team, but only the Core Team will be spotlighted in any complied / published work.
- Participate actively if you want to make the team! Leads consider that.
- Core Team shortlisted. Intensive work begins!
Session #4+ [Tentative length: 1.5 months]:
- Core Team will work towards the core problem.
- Will be recognized by AISC at the conclusion of the project!
- [For your portfolio / blog / Social Media].
- If at the end of the 2 month period, there are new areas we’d like to explore, then we may have a follow-up 2 month Working Group! And then another, and so on.
Aggregate Intellect hosts one of the most diverse ML communities in the world. Over the course of the working group
- You’ll get an immersion into that community & walk out with some cool new friends.
- Learn and get hands-on experience on Forecasting
- Contribute to a repository that will be seen/used by many
- Potentially publish an article on the topic