http://www.sciencemag.org/news/2017/02/artificial-intelligence-grows-nose
22 teams of computer scientists have unveiled a set of algorithms able to predict the odor of different molecules based on their chemical structure. It remains to be seen how broadly useful such programs will be, but one hope is that such algorithms may help fragrancemakers and food producers design new odorants with precisely tailored scents.
This latest smell prediction effort began with a recent study by olfactory researcher Leslie Vosshall and colleagues at The Rockefeller University in New York City, in which 49 volunteers rated the smell of 476 vials of pure odorants. For each one, the volunteers labeled the smell with one of 19 descriptors, including "fish," "garlic," "sweet," or "burnt." They also rated each odor's pleasantness and intensity, creating a massive database of more than 1 million data points for all the odorant molecules in their study.
[...] The upshot is that even though the current study showed computers can predict which of 19 words people will use to describe this set of odors, it's not clear whether the same artificial intelligence programs would rise to the challenge if there were more categories.
Predicting human olfactory perception from chemical features of odor molecules (DOI: 10.1126/science.aal2014) (DX)
(Score: 0) by Anonymous Coward on Tuesday February 21 2017, @02:48PM
You're kidding... this can be done analytically based on spectrometry and a knowledge of human smell receptors. It's like they went out of their way to make it as unscientific as possible.
(Score: 2) by ikanreed on Tuesday February 21 2017, @03:54PM
Yes. AI experiments tend to orient themselves towards solving subjective problems rather than the objective ones. Science, as much as it possibly can, prefers to deal with the objective, because the subjective can't be structured as a falsifiable hypothesis.
Oddly enough, the similarity between two subjective assessments, one from humans, one from a machine, in spite of both sources being subjective, can be tested objectively. And that's precisely what this experiment uses as their dependent variable.
(Score: 0) by Anonymous Coward on Tuesday February 21 2017, @04:42PM
No idea what you are going on about, but from the abstract: "Regularized linear models performed nearly as well as random forest–based ones, with a predictive accuracy that closely approaches a key theoretical limit."
So the original AC is right, there is nothing here except researchers publishing some simple machine learning results in a tabloid to cash in on hype.
(Score: 2) by ikanreed on Tuesday February 21 2017, @04:49PM
Existing technology applied to novel problem class. That's... pretty much all R&D research.
(Score: 0) by Anonymous Coward on Tuesday February 21 2017, @04:51PM
This isn't something to publish a paper about though... it is like something you do as a tutorial to learn an R/python library.
(Score: 0) by Anonymous Coward on Tuesday February 21 2017, @03:43PM
Well that stinks... Lol
(Score: 1, Funny) by Anonymous Coward on Tuesday February 21 2017, @04:16PM
really will be packet sniffers