[FRIAM] Judea Pearl: Book of Why

[Steven A Smith]

> One way to address the N/A issue is to repeatedly perturb the real-world system so as to elicit those correlations.  When that is practical.. 

We are, in a time of real-world system perturbation, right now.  The
whole world is responding to what is *roughly* the same virus with
*roughly* what is the same human phenotype/metabolism in a myriad of
*roughly* the same modes of human organization.   This IS a testbed of
human (-system?) response to a widespread, somewhat invisible threat.  
From Wuhan to Singapore to Italy to Iran to Sweden to Germany to NYC to
WA State to the Navajo Nation to Florida's beaches, this IS a huge
coupled systems dynamics/agent-model executed in real-time by
real-people with real casualties and real consequences.  

We are, to varying degrees (collectively) recording the results of these
"experiments" and if we are lucky (or smart, or both) we will do some
post-game analysis intended to understand more-better how best to
(self-)organize around a (nearly) existential world-scale threat.   And
to the extent this is a game that will never end, we have to begin the
analysis while we cope with it's consequences.   Feels a bit like the
models pof Physics Interreality.

    https://journals.aps.org/pre/abstract/10.1103/PhysRevE.75.057201
<https://journals.aps.org/pre/abstract/10.1103/PhysRevE.75.057201>

Hanging too aggressive of a model on this (or collecting the data 
against too premature of a model) will reduce the utility of such data
gathering and analysis.   Whatever the dual of overfitting a model is? 
Overmodeling?  Premature Modeling?

What I'm looking (askance) to(ward) Pearl for is a better way of rapidly
constructing, maintaining, revising as generic of a model as possible in
response to "this moment".   Four months ago we should have been
interested in models of how one limits a virus such as COVID19 getting a
foothold in this country.   One month ago we should have been interested
in how one limits COVID19 (with new understanding of it's virility, it's
fatality, it's symptoms, it's mode of spread) once it HAS a foothold, 
now we are faced with trying to understand how to cope with it once it
is pervasive in our population whilst continuing/returning to "business
as usual" and in another thread, I'm encouraging that we "try to
plan/consider/think-about" what we want to do with this somewhat "blank
slate" (our ass?) we are having  handed to us.  

And how to think about this without premature modeling... what I think I
was railing (whining/pushing-back) about with Dave on the Bellamyist
thread earlier this morning.

- Steve

>
>> On Apr 19, 2020, at 8:33 AM, uǝlƃ ☣ <gepropella at gmail.com> wrote:
>>
>> Well, the argument I often end up making is that you can do a kind of face validation with the fake data. Show it to someone who's used to dealing with that sort of data and if the fake data looks a lot like the data they normally deal with, then maybe more data-taking isn't necessary. If it looks fake to the "expert", then more data-taking is definitely needed.
>>
>>> On 4/19/20 8:29 AM, Marcus Daniels wrote:
>>> I have a hard time with this as a way to extend data.   If it is high-dimensional it will be under-sampled.  Seems better to me to  measure or simulate more so that the joint distribution can be realistic.  And if you can do that there is no reason to infer the joint distribution because you *have* it. 
>>>
>>>>> On Apr 19, 2020, at 8:18 AM, Frank Wimberly <wimberly3 at gmail.com> wrote:
>>>> 
>>>> Going back and forth:  If you infer the causal graph from observational data you can use that graph to simulate data with the same joint distribution as the original data.
>> -- 
>> ☣ uǝlƃ
>>
>> .-. .- -. -.. --- -- -..-. -.. --- - ... -..-. .- -. -.. -..-. -.. .- ... .... . ...
>> FRIAM Applied Complexity Group listserv
>> Zoom Fridays 9:30a-12p Mtn GMT-6  bit.ly/virtualfriam
>> unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com
>> archives: http://friam.471366.n2.nabble.com/
>> FRIAM-COMIC http://friam-comic.blogspot.com/ 
> .-. .- -. -.. --- -- -..-. -.. --- - ... -..-. .- -. -.. -..-. -.. .- ... .... . ...
> FRIAM Applied Complexity Group listserv
> Zoom Fridays 9:30a-12p Mtn GMT-6  bit.ly/virtualfriam
> unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com
> archives: http://friam.471366.n2.nabble.com/
> FRIAM-COMIC http://friam-comic.blogspot.com/ 
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20200419/afd3dc26/attachment.html>