4 ms·
Yes. I think the author meant that the rate of change is bounded [without the function needing to be differentiable] but used the term "gradient" anyways. I gu
by porridgeraisin 1y ago
Yes.
I think the author meant that the rate of change is bounded [without the function needing to be differentiable] but used the term "gradient" anyways. I guess you shouldn't use the term "rate of change" either? I guess
The Lipschitz property means the function's output doesn't change faster than linearly with input
Is precise in that context
- constantcrying 1y agoHe even gives the definition for the L=1 Lipschitz condition afterwards. Lipschitz continuity means the differential quotient is globally bounded (where it exists). Differentiability, by definition, means the differential quotient converges at every point. These are somewhat related, but different things.