Do non-fallacious cases of argument from ignorance exist?

Some critical thinking authors (i.e., Douglas Walton) contend that there are cases when appealing to the argument from ignorance is reasonable. I'm not fully convinced. If you believe there are, please provide some examples. Thank you!

I am reminded of Sherlock Holmes (AC Doyles) Aphorism. "When you have established what is impossible and exlcuded it, what is left, no matter how implausible, must be the truth". This is a form of deductive reasoning, and in a way this is what the argument from ignorance is trying to do. Though deductive reasoning is useful in things like math, there are a couple of reasons it always seems to fall short when used in reality, especially when used in the form of teh argument from ignorance.

First, deductive reaoning is invalid in cases where there are an infinite or at least unknown number of possibilities. You can subtract away individual possibilities or catagories of possibilities for specific reason but you can not give arguments or have evidence against possibilities you haven't imagined.

I like to say "Absolute certainty is not a triumph of knowlege but a faiure of imagination."

That said, there are two caveats I would add which sometimes make things like an argument from ignorance useful.

One i have alluded to, when your serarch space is well defined and finite, as it is in some mathemtical systems, you can use the faiure of other explantions to bolster the probability of remaining cases being true.

Secondly, there is a bayesian probabilitic version of it that in some cases might be valid. When you are familiar with the the likely things hat might be true, if you can eliminate all but one then you can ignore the vanishingly samll probabilities you have not considered and assume what remians is probably right.

Example, lets say that a car mechanic has fixed 500 toyotas that would not start. He has found that 350 of them had bad ignitions, and 150 had bad fuel pumps. If he tests the ignition on car 501 and find it puts out a good spark, its is reasonable for him to assume that it is highly probable that car 501 has a bad fuel pump. This only works because he has prior knowlege of probabilities in the system he is working with, and because he is making a practically probabilistic statment, not claiming a logical absolute.

You can't use argumenst like this the way Argumenst from ignorance frequently are used, to demonstrate teh existence of singular or previous unproven phenomena. If someone sees something he is sure is a dog, and claims it is a dog because we know people rarly misidentify dogs, we have enough experience with people seeing dogs that turned outto be dogs to trust his assertion. On the otherhand, if someone sees a UFO and is certain that it is an aliean space creaft because he can think of no other explanation, we can draw no conclusions about whether his perception and conclusion are reliable because we have no prior experienc with detrmining whether people can reliably identify aliean space-craft versus all other possibl UFO's.