Join NCI experts to discover how deep learning is reshaping weather and climate science, and gain the skills you need to start exploring these powerful approaches in your own work. This online workshop includes a public webinar, followed by optional on-demand hands-on sessions (see "Format" section for details).

The webinar session will provide an overview of recent developments in the field, showcasing several state-of-the-art weather and climate models and the resources available at NCI to support the Australian research community in advancing deep learning applications and scientific understanding.

Participants will also be introduced to the NCI Inference Server, which enables skill comparisons across models and performance benchmarking. The session offers an opportunity to provide feedback on example models, suggest improvements, and share your needs for future support. We also invite you to tell us which models or algorithms you are currently using or interested in exploring.

The hands-on sessions (prerequisites apply) will take a deeper dive to the technical details using Jupyter notebooks. Tailored for environmental data, these sessions will explore example network architectures, training and inference strategies used in example models.

Format

You can choose to attend the webinar only.

You must attend the webinar in order to get the most out of the hands-on workshop.

30 October 2025	1 pm - 4 pm (AEDT)	Webinar	All welcome
6 November 2025	1 pm - 3:30 pm (AEDT)	Hands-On Session	Check prerequisites below

Who Should Attend

Webinar

The webinar content is developed with weather and climate researchers in mind but anyone who is curious about the topic is welcome.

Hands-on Sessions

People using or interested in training, fine-tuning or using deep learning models for weather and climate problems, including but not limited to:

Weather and climate researchers and practitioners
Research software engineers
Higher Degree Research students

Prerequisites

Basic background in climate and weather science.
Familiar with deep learning concepts.
Have Python programming for scientific data analysis experience.
Attended the webinar session

Agenda

Thursday 30 October 2025, AEDT	Webinar
1:05 pm - 1:10 pm	Dr. Rui Yang	Opening
1:10 pm - 1:30 pm	Dr. Yue Sun	A Gentle Introduction to Deep Learning Models
1:30 pm - 2:00 pm	Dr. Maruf Ahmed (Hands-on session available for these models)	AIFS-single AIFS-CRPS GenCast FourCastNet3
2:00 pm - 2:15 pm	Dr. Edison Guo	BARRADiff
2:15 pm - 2:30 pm	Tea Break
2:30 pm - 3:00 pm	Dr. Yue Sun (Hands-on session available for these models)	LUCIE: a lightweight uncoupled climate emulator costs <500SU to train. Aardvark Weather: an end-to-end model which takes in observations from satellites, weather stations and weather balloons, and produces a ten-day global forecast.
3:00 pm - 3:30 pm	Dr. Rui Yang	More NCI Support Inference server to compare skills across models Performance Benchmarks
3:30 pm - 4:00 pm	Dr. Yue Sun	Discussions Feedback on the example models. Suggestions to further improvements. What more support do you need? What models/algorithms are you interested in / working on?
Thursday 6 November 2025, AEDT	Hands-On Session
1 pm - 1:30 pm	Zhuochen Wu	Cluster access setup using ARE
1:30 pm - 3 pm	Dr. Yue Sun	LUCIE Data Inspection Replicate Training Modified Training Inference Aardvark Weather Data Inspection Model Forecast Finetuning Encoder Training
3 pm - 3:30 pm	Dr. Maruf Ahmed	AIFS-Single-v1.1 Demo

Recordings

This event will NOT be recorded.

Contact the host of Deep Learning Model Development in Weather and Climate Studies

There were 0 errors found in the information you submitted: