nostalgebraist reblogged this from lostpuntinentofalantis and added:

What immediately came to my mind is that the Bayes setup doesn’t demand that your prior be continuous in any underlying...

actual-mandible-garrus-labeouf liked this

locally-normal liked this

lostpuntinentofalantis reblogged this from locally-normal and added:

I am a bad pun blog and I endorse this message as elaborating on my “eh it probably converges” intuition earlier. I...

jack-rustier reblogged this from nostalgebraist

locally-normal reblogged this from nostalgebraist and added:

This sounds like a natural continuity/limits problem. It does seem like there could be infinite nesting like this, and...

jadagul liked this

der-unverantwortliche liked this

argumate liked this

kitpeddler liked this

humpolec liked this

type12error liked this

nostalgebraist said: (the follow-up posts where jadagul framed things in terms of uniform vs. pointwise convergence, and i riffed on that, are worth reading. because the convergence is pointwise, some properties of the limit may *never* appear at any finite time [as in gibbs phenomenon], which calls into question the idea that the LI is really imitating the limit in the way one might intuitively want.)

nostalgebraist said: @derplefurf: i lose you at “Analogously, this shows…” the LI criterion can be satisfied computably (LIA), but i argued that you can satisfy it and still be terrible at inference. before we even talk about efficient computation, we need to establish that we are approximating a good thing, and the LI criterion doesn’t get us that.