AI can renew a vision that was once doomed to fail

When you make movies Leaping from silence to sound, production companies were eyeing the next sensory frontier in entertainment.

So in 1939, they tried to smell the cinematic experience. The Aroma-O-VisionIt is a system that transfers the packaged scents from under the tubes movie theater Benches, debuted at the 1939 World’s Fair in New York City.

It didn’t go as planned. The public complained that Aromatherapy It was out of sync with the movie, overbearing, or simply unpleasant. Although a revival was attempted in the 1960s, the technology has largely fallen out of favor.

But now, more than sixty years later, the science is a lot closer to making Smell-O-Vision a reality.

The Smell-O-Vision concept didn’t work as planned, but it may be due to a revamped AI.LMPC / LMPC / Getty Images

Computer scientists and chemical engineers at the Tokyo Institute of Technology in Japan have developed a machine learning An algorithm that can reverse engineer an odor based on its chemical composition.

With this technology, they hope to one day create custom scents on demand, according to a study recently published in PLUS ONE. The smell of ratatouille seems tempting as it is carefully prepared ratatouille.

Here’s the background – Smell played an important evolutionary role: in the brain The olfactory bulb, which is responsible for processing smells, is located next to the amygdala that deals with emotions – and the two share a significant neural overlap.

That’s why some scents tend to elicit a strong emotional response; For example, the smell of chocolate chips may transport you to the warmth of your grandmother’s kitchen, while the smell of pencil shavings may trigger college exam anxiety.

We rely on these subtle olfactory cues to tell us whether we feel safe, stressed, excited, or relaxed in a particular environment.

The olfactory bulb, which processes smells, is located next to the amygdala, which processes emotions.TIM VERNON / SCIENCE PHOTO LIBRARY / Science Photo Library / Getty Images

People differ in how they interpret these signals. Factors such as an individual’s race, sexGenetics and unique experiences can all contribute to the way they perceive a particular scent.

Amanda HollomanA doctoral student in computer science at the University of Alabama who specializes in olfactory research, noted this subjectivity in her work (she was not involved in the new study). “One person might smell popcornAnd another scent smells like vanilla.”

what’s new – The researchers ran each scent through a mass spectrometer to reveal its chemical signature. They were able to isolate about sixty components called odors that make up any scent.

Then, using a machine learning algorithm, they set a quality – like sweet, fruity, or astringent – for each smell, and measure its ratio in relation to the other components of the smell. The researchers say that by adjusting these ratios, they may eventually be able to produce personalized scents.

It says “Currently, there is only numeric arithmetic” Takamichi Nakamoto, an engineer at the Tokyo Institute of Technology and first author of the new study. But “later, we’d like to replicate that actual scent.”

Future VR experiences could include custom scents for more immersion for users.Liyao Xie / Moment / Getty Images

why does it matter – The team envisions an “AI-powered perfume maker” that can mix a scent based on a given description. Such a device will be a boon for the cosmetic perfumery industry – for example, cooking new aromas for shampoos, lotions or candles.

But it can also have medical applications, such as Treating certain types of seizures. It can also deliver more realistic virtual reality experiences in the Metaverse and other digital environments.

What’s Next – The algorithm created by Nakamoto’s team was trained to recognize odors using human descriptions of odors from the database.

But because olfactory computing is a relatively new field, Nakamoto says we have limited data on how people perceive odors. This can lead to algorithm bias, which is an increasingly common problem in Artificial intelligence.

To avoid this bias from intrusion – or at least reduce it – Holloman believes that researchers should develop a database using information from people with as wide a range of different backgrounds as possible. “We need to focus more on employment,” she says.

For now, she believes, the new study represents a promising step forward in the field of olfactory computing.

It remains to be seen whether or not Smell-O-Vision will work (or rather, it has been sniffed). Either way, using the power of fragrance with AI sounds pretty cool.