Week 4 — Foundation Models in Research

Held in winter 2026–27 — dates TBD. Prerequisite: Weeks 1 and 3.

Week 4 builds on the deep learning core from Week 3 and extends it into research practice. The goal: by the end of the week you can evaluate models rigorously, build on foundation and other pre-trained models, and incorporate them intelligibly into your own research — with the discernment to know when they help and when simpler methods win.

Objectives

By the end of the week you should be able to:

• Evaluate deep learning and foundation models rigorously — benchmarks, model cards, validation, and verification.
• Understand the foundation model landscape for Earth systems science and pull, run, and adapt models for your own scientific problems.
• Build on other models — from prompting and embeddings through fine-tuning — rather than training from scratch.
• Integrate models intelligibly into research pipelines and agentic systems.
• Judge when a foundation model is the right tool, use models to inform mechanism, and communicate results and limitations to non-technical stakeholders.

The week at a glance

The day-by-day schedule will be posted when the winter dates are announced.

The Week's core:

Using foundation models is a direct way to stand on the shoulders of many giants.