On the “getting real world stuff” done outside of academia, there’s IARPA’s project of metaphor analysis (for spying programs): Last I checked, not only did they get a winning bid, but they constructed the necessary AI software for them and that phase of the project is completed.
Good thing about ARPA is eventually ‘the people’ get to benefit from many of their programs. [see: intergalactic computer network]
Isn’t that OWL? Gotta research a little. I *think* that’s more related to what became semantic web ontologies like the ones from Web 2.0 days…
I’m not sure, but I think that project was ground up. They wanted to analyze cultural metaphors real-time, for middle-eastern intelligence, creating models of potential meaning in order to code-break artificially constructed metaphors based upon already known cultural metaphors that might be invisible to people raised in the USA.
An example from internet culture would be attempting to understand a meme like this using neural networks programmed with cultural references to properly interpret it.
Or, even harder, something like this. Now, these are image metaphors and not written/verbal metaphors but imagine if you were certain, due to the actors communicating, that this meme contained critical information that could help prevent a terrorist attack.
Indeed. Feature identification is getting better and better. Had I tried I would’ve done first to tineye which has a fantastic image matching service, able to catch broad feature similarities.
I’d then go through, starting with the older references and also ones from ‘decent’ sites, to see if there’s an explanation of it. [usually there is somewhere].
After _that_, I’d probably break the features down myself and search again.
Wait: scratch all that: I’d wait ’til my 11 yr old nephew got up tomorrow and ask him. “What’s this face mean?” and he’d tell me.
Or an easier method: This says “ifunny.co “. I have iFunny on my phone. So I’d find the image there and look in the comments
Reminds me of a certain animation style I never cared for from the 80s or so