@bjonnh I was playing with new symbol API and I managed to befriend it with autodiff and commons math optimization. It is still very preliminary since I did not manage to make commons-math optimization safe enough, but the example from this test works fine both with simplex and gradient descent.

This is cool

A cleaned-up API: https://github.com/mipt-npm/kmath/blob/feature/optimization/kmath-commons/src/test/kotlin/kscience/kmath/commons/optimization/OptimizeTest.kt. It is not final on the inside but close to final on the outside. Again, comments and criticism would be appreciated. The final step will be to add X-Y data fitting.

Pretty busy this week, but I want to give it a try

I've finished an example with x-y fit. It looks really nice since it combines both autodiff and gradient descent in one place with full customization, but the design is not final. There are still things I find non-intuitive. The example is the last test in the reference above.