Taxacom: An artifical 'intelligence' and key

[John Grehan]

Some decades ago back in the 1990's one of my supervisors wanted to look
into the application of 'fuzzy logic' to the identification of invertebrate
samples collected during biodiversity surveys. I think he had in mind the
challenge of dealing with massive quantities of specimens across the
taxonomic spectrum. He conceived the possibility that with comparative
identification photos, one might photograph each specimen and leave it to
the algorithm to at least narrow down possibilities. As this could only
come through 'competitive' funding, and support from appropriate
algorithm specialists, nothing came of it. I just looked up the 'fuzzy
logic' term on the web and the first entry was " a type of artificial
intelligence (AI) that uses fuzzy sets to solve problems. They are designed
to mimic human decision-making by considering all possible outcomes between
"true" and "false"

On Sun, Feb 9, 2025 at 8:27 PM Peter Uetz via Taxacom <taxacom at lists.ku.edu>
wrote:

>
> The question is whether they should. Dichotomous keys are (usually) a pain
> to use, and if you have one less reliable character, they may completely
> mislead you.
>
> However, LLMs can produce character tables from descriptions, which you
> can use for ID purposes. The output, of course, depends on your input. Here
> are some examples (including good and bad ones):
>
> Uetz, P.; Darko, Y.A.; Voss, O. 2023
> Towards digital descriptions of all extant reptile species.
> Megataxa 010 (1): 027–042
> https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fdoi.org%2F10.11646%2Fmegataxa.10.1.6&data=05%7C02%7Ctaxacom%40lists.ku.edu%7C320d72f575644b2ce2a108dd4a19c7c7%7C3c176536afe643f5b96636feabbe3c1a%7C0%7C0%7C638748196772536015%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=7sH%2BVoBjpxDCpVKZECYG2UVfJeX2QOnS6IJjKT0ewow%3D&reserved=0
>
> Let me know if anybody has tried this for other groups, including insects!
>
>
> > On Feb 9, 2025, at 19:11, Vr.R.Bejsak-Colloredo-Mansfeld via Taxacom <
> taxacom at lists.ku.edu> wrote:
> >
> > Hi
> > Just wonder when a software with Artificial inteligence will be able to
> create identification key?
> >
> > Regards
> >
> > Vratislav
> > Sydney
> > retired entomoogists
> > _______________________________________________
> > Taxacom Mailing List
> >
>
> ————————————————————————————————————
> Peter Uetz
> Center for Biological Data Science
> Virginia Commonwealth University
> Richmond, VA 23284
> USA
>
> Reptile Database
>
> _______________________________________________
> Taxacom Mailing List
>
> Nurturing nuance while assailing ambiguity for 38 years, 1987-2025.
>
> Send Taxacom mailing list submissions to: taxacom at lists.ku.edu
> For list information; to subscribe or unsubscribe, visit:
> https://lists.ku.edu/listinfo/taxacom
> You can reach the person managing the list at: taxacom-owner at lists.ku.edu
>
>
>

-- 
https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fhepialidsoftheworld.com.au%2F&data=05%7C02%7Ctaxacom%40lists.ku.edu%7C320d72f575644b2ce2a108dd4a19c7c7%7C3c176536afe643f5b96636feabbe3c1a%7C0%7C0%7C638748196772553729%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=Kf3%2Bfqb7gRESRRwGILNtLJFWiHInSdKzDjhp8GpUsoA%3D&reserved=0 (use the 'visit archived web site'
link, then the 'Ghost Moth Research page' link.