Automatic differentiation in scientific programming with jax

We use derivatives extensively in science and engineering. Historically derivatives have been a challenge in computer programs. They can be tedious to derive and implement, or difficult to approximate. Today a new option exists called automatic differentiation, where a computer program can take the derivative of a program, automatically. This alleviates the need to derive and implement a derivative. It is equivalent to an analytical derivative, and more accurate than an approximation. This is a paradigm changing development in scientific programming. It is now possible to write programs with derivatives easily. The jax library in Python provides a modern implementation of automatic differentiation that is often used in machine learning, but which can be used in many other ways in pycse. In "Automatic differentiation in scientific programming with jax" you will learn how to use jax to solve many kinds of math, science and engineering problems.
You can find the booklet at pointbreezepub...

[@youngsci]

Thanks you very much for this amazing video

[@men6829]

Thank you very much !

[@TraderAli]

Thanks for wonderful work, at 2.30, using np does not work for me, instead import jax.numpy as jnp and using it, works

[@JohnKitchin]

I almost always import jax.numpy as np. Most likely that had been done earlier. You are right that using regular numpy doesn't work there.