How to fit any dataset with a single parameter

Re:How to fit any dataset with a single parameter by mgraczyk on Wednesday September 29, 2021 @04:47PM (#28699705)

I wouldn't say that the distinction is "unbridgeable". Keep in mind that we send packets over plenty of "geometrical" and seemingly infinitely dense analog channels all the time, like electricity over a copper wire or EM waves over the air.

The "mark on a stick" channel has a capacity like any other channel. If you're sending just one symbol, you could easily calculate the information capacity given a desired probability of bit-error.

Assuming you can put the mark in exactly the right spot, you can model the "noise" as a distribution over the values that the reader will measure. If you model this as `mark + zero-mean normal distribution` with a known variance, then your stick is just an AWGN channel.

[ Parent | on HN ]

Re:How to fit any dataset with a single parameter by dtagames on Thursday September 30, 2021 @02:35AM (#28703771)

The gentleman is not suggesting that quantized information cannot be fully represented in a continuous function, but the opposite. That the continuous functions of life (his muscle example) and consciousness (AI example) cannot be simulated in any meaningful way by a discrete nomenclature or algorithm. The idea is impossible by definition and can be dismissed immediately.

[ Parent | on HN ]