All guide videos

These videos go with the written lessons. The lessons have every detail; the videos show the same steps on screen. Narration is in English with an AI voice. Each video also has a link to the lesson that goes with it.

1. What is scientific machine learning?

Scientific machine learning (SciML) combines machine learning with scientific knowledge. A model can learn from data and also follow known physical laws, such as the heat equation. This video explains the idea with a simple example, estimating temperatures across a metal plate from only a few sensors, and names four common approaches: physics-informed neural networks, neural operators, learning unknown physics, and fast surrogate models. It also reminds you that a SciML model still has to be checked against experiments or a trusted numerical solution.

Next: What the Studio is, and where your work lives

2. Using the Studio or coding it yourself

A short reel that puts two ways of working next to each other: building a scientific machine learning model visually in the Studio, and writing all of the code by hand. In the Studio you drag blocks onto a canvas, connect them, copy the generated code and run it. Watch it to see how the two approaches differ, then decide which one suits your project.

Next: Find your way around the studio

3. Start your first notebook

This video shows how to build a physics-informed neural network (PINN) in Scientific ML Studio. It needs no installation and runs in your browser, and your data is not uploaded. You start a notebook on an empty canvas, drag in each block from the palette, connect the blocks with wires, run the Graph check, and generate the Python code for the network.

Follow along with the written lessons:

• Create a notebook
• Build a notebook from an empty canvas
• Check the graph and generate the code