[FRIAM] heuristics for fuzzy objectives (was Re: Drones to detect wildfires)

[uǝlƃ ☤>$]

Another context in which a *softening* of objectives, or perhaps multi-objective orientation, might help:

Safety Third
https://www.jems.com/operations/safety-is-third-not-first-and-we-all-know-it-should-be/

This context does a better job of highlighting the point that objective-less wandering is *not* unequivocally better than [multi-]objective optimization. Of course, Stanley glosses this in his comments about "modest" objectives. Perhaps we could reframe the idea that an emergency responder's desire to both help someone and not be injured in the process is a "modest" [set of] objective[s]? But it should be easy to see why that's dodgy rhetoric. Staying alive while helping people is "modest"? Really?

The "safety first" mantra is not, unlike the strawman set forth in this article, a hard *ordering* of priorities. The mantra is an attempt to get the individual to think dynamically, adaptively. We know your dopamine-fueled risk-taking personality will *cause* you to leap into the fray before engaging your conscious mind. So our (perhaps flawed) attempt, here, is to encourage you to keep your conscious mind engaged and *balance* all the rational priorities you know apply.

Stanley's rhetoric was hyperbolic if applied literally, especially to situations in the middle of the spectrum between "modest" and "ambitious". But given the uphill battle, it was useful hyperbole. It was a great shock-jock technique. An "analog" to clickbait. Hook 'em quick. But 5-6 years after peak rhetoric, we should be able to couch it in a reasonable spectrum. Heuristics for when to formulate and solve for crisp objectives versus when to *play* versus a mix of both seem more interesting at this point.


On 5/28/21 10:06 AM, Steve Smith wrote:
> The most fundamental problem with the general idea of setting objectives and then using any number of clever methods to converge on those objectives is that *at best* they are only as good as the objectives themselves.   The point being made here, of reducing the granularity to a pieceweiz, forward chaining is well motivated IMO, as it seems the "best we know how to do".
> 
> In the spirit of "exaptation", and "life is what happens while you are making other plans",  the point of picking an objective and searching for it on an exotic, high dimensional (even fractal) landscape is to provide a heuristic for *exploring* the parts of the landscape you can't even imagine much less predict exists.   I would compare it more to "creative wandering" than "active seeking", though I think the two can be compatible.
> 
>     /“It is good to have an end to journey toward; but it is the journey that matters, in the end.”///
> 
>         ― Ursula K. Le Guin, The Left Hand of Darkness <https://www.goodreads.com/work/quotes/817527>

-- 
☤>$ uǝlƃ