Just like a fruit fly, a new algorithm never forgets old scents
Ars Technica Federica Sgorbissa
A new algorithm tries to copy how fruit flies remember smells. It may help electronic noses stop forgetting the old scent the moment a new one shows up.
Based on reporting by Ars Technica, Federica Sgorbissa — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Fruit flies are tiny, but their noses are stubbornly good. With about 140,000 neurons — a brain smaller than a poppy seed — Drosophila can pick apart a wide range of smells in a split second and still remember them later. That combination has made them an awkward benchmark for human-made scent detectors, which tend to be pricier, narrower in what they can sense, and quick to lose track of an odor once they learn another one.
Kevin Max and Yang Shen at the Okinawa Institute of Science and Technology took that problem seriously enough to build an algorithm around it. Their system, called Spi-Fly, is described in a paper recently published in Neuromorphic Computing and Engineering. The idea is simple enough to say and harder to do: make machine smell behave a little more like a fly’s.
The appeal here isn’t that fruit flies are cute lab mascots. It’s that they manage something electronics still struggle with — fast recognition without immediately wiping the slate clean. Spi-Fly is aimed at that exact weakness, the annoying habit of current electronic noses to forget old odors as soon as new ones arrive.
That matters because scent detection only helps if the detector can keep its bearings over time. A device that is impressive for one narrow task, then gets confused the next time the air changes, is not much use outside a demo. The fly, rude little creature that it is, apparently has the better design.
My take — AI-written commentary, not fact-checked reporting
The real story is not that nature is clever; it’s that our fancy machines still forget like goldfish. Spi-Fly sounds like the kind of borrowed biology that actually deserves attention, because it targets a practical flaw instead of polishing another lab demo. The industry loves sleek sensors, but memory is the unglamorous feature that keeps getting ignored.
Read more about this at: Ars Technica