Long and I spent part of the morning arguing about whether our model for the insane robot astronomical censored catalog data stream is mathematically consistent. We decided it is (and then I went of for an awesome afternoon at the Exploratorium). It helped that we made this graphical model (typeset in LaTeX by Foreman-Mackey):
There are variables with no parents inside plates (A_k, B_k, and t_i). My mental graphical model lint tool is suspicious. Are you sure you know the marginal distributions of those variables? From multiple observations, one could potentially learn their distributions, but you're *insisting* you already know them.
ReplyDelete(Of course I don't know what those variables are, so it might be perfectly fine.)
The principles behind this, a sort of Maxwell's daemon in the entropy of it all with sentience- well, if your forgive my lack of reductionism delving into the philosophy, it add up but one should question the sanity of it anyway. My posts at pesla.blogspot suggests what I think is similar to this in the metaphysics of it. I doubt it can be reduced into simply quantum terms. The series I tentatively call Overworlds.
ReplyDeleteNow, you are quite interesting in your posts to me for your ideas in the practical observation and interpretations seem on the brave frontier- I hope someone is listening. I had a thought about you today for hidden things such as data in in moves through the next dimension- that is your mapping of the fractal as a plane filler--- that if we used the other one it would be a wider method of computing errors and resolution of this unseen presumably hidden data.
What counts I suspect is what works better even if it is foundational super uncertain philosophy. Your are on to something, thank you... it most certainly is the right direction. But I am not trained at all in seeing this.
Poetry said Ferlinghetti in the end is about cosmology.
The PeSla