Your next signature scent might be designed by a robot

Widely considered the most mysterious of the five senses, scent is uniquely personal; that’s why the iconic Chanel No. 5 smells like the epitome of elegance to some, and cat urine to others. Now, artificial intelligence technology developed by IBM can offer some insight into consumers’ olfactory preferences, using machine learning to fill gaps in the perfume market.

In an interview for a new book by gestalten on the history, science and culture of the perfume industry, The Essence – Discovering the World of Scent, Perfume & Fragrance, IBM research scientist Dr. Richard Goodwin discusses developing and working with the algorithm-based artificial intelligence tool Philyra. Goodwin says that the technology can be used to create new perfumes catered specifically to different age and gender demographics. Who could have guessed that Brazilian men aged 25 – 40 tend to enjoy the smell of cardamom and carrot seeds?

Though the technology isn’t perfect (don’t expect to see human scent experts phased out in the near future – German fragrance company Symrise still employ master perfumers to tweak the Philyra-generated formulas to get them just right) this marks a shift in the world of product development, where AI like Philyra can now take the place of focus groups and streamline market research. After all, if companies can optimise something as enigmatic and subjective as fragrance, the sky is the limit. It’s only a matter of time before AI technology becomes standard in developing ultra-personalised cosmetics and skincare.

Philyra is an AI tool (developed by IBM Research and Symrise) that uses machine learning to trawl a database of perfumes, spitting out formulas that resonate with particular demographics, as well as conjuring up totally new combinations. IBM research scientist Dr. Richard Goodwin discusses working with Philyra to develop the first-ever AI-designed fragrances (Egeo On Me and Egeo On You) targeted towards millennials in Brazil below. To read more, buy The Essence here.

Photo courtesy of Gestalten

Using AI to create a fragrance sounds very sci-fi – how does Philyra work?

It uses new and advanced machine learning algorithms to sift through hundreds of thousands of fragrance formulas and thousands of raw materials to help identify patterns and novel combinations. So Philyra does more than serve up inspiration – it can design entirely new fragrance formulas by exploring the entire landscape of fragrance to discover the ​“white spaces” in the global fragrance market.

Will this take the traditional ​“nose” out of the equation?

No, fine fragrance is an art and a science that takes people 10 years of training to master. Philyra is more like an apprentice, but one that can’t smell, so AI won’t replace master perfumers. Philyra can look at the history of perfumery and learn from the formulas created by perfumers over many decades, but it still needs the expertise of the perfumer.

This year Philyra designed two new perfumes targeting millennials living in Brazil. What type of scent did Philyra create for them?

The perfume for the ​“millennial man” features scents of fenugreek seeds, green cardamom pods, carrot seeds, and more, wrapped in a milky, buttery, rich base. The one for the ​“millennial woman” is fruity and oral and has scents of osmanthus tea with lychee and patchouli.

Symrise used Philyra to design the two perfumes for O Boticário, one of the top global beauty companies. The initial formulas suggested by the system were tweaked by a master perfumer to emphasise a certain note and improve how long it lasted on the skin. Philyra’s understanding of consumer preferences means perfumers can focus on perfecting the final perfumes, rather than spending time searching for new fragrance combinations. Meanwhile, O Boticário could utilise AI to create a perfume tailored to a specific demographic and personality.