[FRIAM] signal and noise

[Marcus Daniels]

A common objection to deep generative learning is that such models don’t offer explanations, at least without clever formulation.  It seems to me that addresses Dave’s concern.  As long as a distribution of measurable things is captured at high fidelity there are no preconceived biases about what matters or what “acceptable” models look like, algorithmically.

> On Sep 9, 2022, at 7:27 AM, glen <gepropella at gmail.com> wrote:
> 
> I think it's reasonable. But I also think it leans wrong, depending on what you mean by "several conversations", "algorithmic", and "noise".
> 
> Marcus' suggestion that there's an irreducible limit somewhere below whatever SnR threshold is recognized is only a *bottom*. The distance between the recognized threshold and the incompressible kernel of noise is non-zero, almost by definition. And Frank's suggestion that there are established methods to tease more signal from that non-zero band, indicates he sees it as well. Hearken back to my and SteveS' discussion of interstitial spaces being dual to the entity-objects they house and you could see us agreeing with you, there, too. Jon and SteveG's discussions of duality tend to be less prosaic, but nonetheless a bit mystical. Contra-pose the back and forth of Nick and EricC's constant assertion of behaviorism, Frank's objection, yet the subtle differences and challenges between them, and it should be clear there's a non-zero band between recognized noise and the incompressible limit. Jochen posts more questions than answers. Even EricS' conversations with Jon about the expressive power of hypergraphs shows an impetus to circumscribe what's computable and what's not. I mentioned a Wolpert paper awhile back, wherein he gives some air to hypercomputation, to which nobody on the list responded. And you've even defended brute force computation by highlighting the progress and efficacy of techniques like Monte-Carlo simulation.
> 
> I'm sure there are other arguments I've missed. Perhaps you're doing a bit of "othering" in thinking your focus on the noise is unique? But perhaps, given that we're 99% male and *old*, there's a tendency for most of us to pretend to know more than we know? ... to inflate the epistemic status of our pet hypotheses? Humility is punished in most contexts, despite the lip service we pay to it.
> 
> What I see is a persistent inability to play the games set up by others ... an insistence that others always play our own game. When others don't play the game proposed by someone, that someone takes their marbles and leaves. The voyeuristic lurkers may enjoy watching different games, but won't play. Some may be frustrated that some games have no clear rules by which to "win" (i.e. come to a Peircian convergence, a belief in Modernist true Truth). Etc.
> 
> IDK. Here's a paper coming up quick in my queue that may help demonstrate you're not alone:
> 
> The Experimental Evidence for Parapsychological Phenomena: A Review
> https://thothermes.com/wp-content/uploads/2022/04/Cardena.pdf
> 
> And I mentioned a long while back Broderick and Goertzel's similar effort:
> 
> Evidence for Psi: Thirteen Empirical Research Reports
> https://bookshop.org/books/evidence-for-psi-thirteen-empirical-research-reports/9780786478286
> 
> which, again, got no response on the list. No response or hostile response doesn't mean you're unique in your perception or perspective. It can mean many things. The only thing we *might* control is our own attitude. We can choose to see ourselves in those around us. Or we can other those around us and think we're alone. I try to choose the former.
> 
> 
>> On 9/8/22 16:12, Prof David West wrote:
>> It seems, to me, that several conversations here—AI, hallucinogens, consciousness, participant observation, and epistemology—have a common aspect: a body of "data" and disagreement over which subset should be attended to (Signal) and that which is irrelevant (Noise).
>> Arguments for sorting/categorization would include: lack of a Peircian convergence/consensus; inability to propose proper experiments; anecdotal versus systematic collection; an absolute conviction that everything is algorithmic and, even if the algorithm has yet to be discerned, it, ultimately, must be; etc..
>> I often feel as if my positions on these various topics reduces, in some sense, to a conviction that there is overlooked Signal in everyone else's Noise; even to the point of believing the Noise IS the Signal.
>> Is this in any way a "fair' or "reasonable" analysis?
> 
> 
> -- 
> ꙮ Mɥǝu ǝlǝdɥɐuʇs ɟᴉƃɥʇ' ʇɥǝ ƃɹɐss snɟɟǝɹs˙ ꙮ
> -. --- - / ...- .- .-.. .. -.. / -- --- .-. ... . / -.-. --- -.. .
> FRIAM Applied Complexity Group listserv
> Fridays 9a-12p Friday St. Johns Cafe   /   Thursdays 9a-12p Zoom https://bit.ly/virtualfriam
> to (un)subscribe http://redfish.com/mailman/listinfo/friam_redfish.com
> FRIAM-COMIC http://friam-comic.blogspot.com/
> archives:  5/2017 thru present https://redfish.com/pipermail/friam_redfish.com/
> 1/2003 thru 6/2021  http://friam.383.s1.nabble.com/