Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

That may not be what you meant but it is what you said. The "function you are trying to model" does not have to be differentiable as we have already established. Only the model itself (which is the neural network) needs to be differentiable.


The model needs to be differentiable.

The function needs to be continuous.

Discontinuous functions are not differentiable.

You can approximate discontinuous functions with continuous ones (e.g. with logistic functions).

In the end, we are still talking about curve fitting and optimization, not artificial general intelligence.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: