Hacker News new | ask | show | jobs
by 0xBABAD00C 1148 days ago
> Gradients Are Not All You Need

Sometimes you need to peek at the Hessian.

Seriously though, what is intelligence if not creative unrolling of the first few terms of the Taylor expansion?

1 comments

Of what?
`a + ax + ax^2+...`
That's the expansion. What does it represent?
Given the context of the original statement, the Taylor series expansion represents the statistical learning process of any topic anywhere.
I think you mean loss function, not process. And I'm failing to make any connection between intelligence and polynomial approximations to loss functions.
Intelligence is rather ill defined which I suppose is why you are having difficulties.
Yes.

(And it goes down all the way)

No that's turtles.