Neural networks and guinea pigs: A credible use case for AI is finally here

Despite the conjecture spouted about the usefulness of artificial intelligence (AI), what’s been sorely lacking (as far as Downtime is concerned) is that one killer use case that really captures just how transformative an impact this technology will have on all our lives.

That was until Downtime came across the ground-breaking work of research scientist, Janelle Shane, who is using neural networks to help animal rescues in the United States pick names for the huge numbers of unwanted pets they take in every week.

Shane’s foray into this area began with an S.O.S from Portland Guinea Pig Rescue, who asked if she would consider using her neural network expertise to come up with credible and authentic-sounding names for all the furballs they take in.

The rescue drew up a list (featuring the name of every guinea pig that has ever darkened its doorstep) to train Shane’s neural network with, and – within a day – the first tranche of cavies with computer-generated names were listed for adoption on its website.

At the time of writing, many of them were still looking for their forever homes, but Downtime is confident they won’t be waiting for long, sporting cool neural network-created names like, Stargoon, Rockass and After Pie.

Two guinea pigs

Fleury White & Stargoon: Two adoptable guinea pigs at Portland Rescue, named using neural networks.

As is often the case with AI endeavours, some of the names thrown back by the neural network have been passed over for missing the mark a little, with “Boooy”, “Bho8otteeddeeceul” and “Fleshy” among Downtime’s favourites to land on the reject pile.

Since the Portland project, Shane’s has since been called upon to repeat the experiment for another guinea pig rescue centre, as well as cat rehoming centre, suggesting we’ve not even scratch the surface of just how far this application of the technology could go.