Symrise and IBM collaborate with O Boticario to deliver First AI Fragrance in 2019

IBM Research has been working with Symrise —one of the world’s leading suppliers of fragrances, to create an AI (Artificial Intelligence) system for creating new scents.

O Boticario’s Heloisa Souza, Symrise’s Achim Daub, IBM’s Principal Research Scientist Richard T. Goodwin, and Symrise Perfumer David Apel

Internally called Philyra, this solution has been built for Symrise using IBM Research’s AI for Product Composition technology and used to create two new fragrances for their client O Boticario which will launch in 2019.

“The art and science of designing a winning perfume has been something we at Symrise have been doing passionately for more than 200 years. Now, our perfumers can work with an AI apprentice by their side that can analyze thousands of formulas and historical data to identify patterns and predict novel combinations, helping to push the boundaries in perfumery, and accelerate the design process by guiding them toward formulas that have never been seen before,” stated Achim Daub, President of Symrise Scent & Care.

“This effort brings together human expertise and sensibility with machine intelligence to create novel, fine fragrances. Philyra’s understanding of consumer preferences and knowledge of formulas and ingredients, led to new fragrance combinations which allowed our perfumers to accelerate the creative design process and focus on perfecting the final perfumes — the alchemy that helps us address our consumers’ needs in a unique way,” said Alexandre Bouza, Marketing Director, O Boticario. “Advances in AI are helping us introduce unique, new fragrances to our customers. We are very excited to be part of this collaboration with Symrise and IBM Research.”

O Boticario’s Tiago Martinello, and George Ledes

Operating in a market that is estimated to be approximately $17.4 billion in size by 2024, Symrise’s Scent & Care Division creates and sells fragrances and aroma molecules used in fine fragrances, personal care and beauty products, as well as home care
products such as for cleaning, hair care and detergents.

In the world of fragrance design, fragrances are traditionally created by perfumers who train for many years before they become proficient at their craft in developing a winning fragrance that is consumer preferred, triggers the desired emotional response, and is considered unique and innovative.

As part of a multi-year-long effort, Symrise and IBM Research formed a collaboration to explore the application of AI to help create new and innovative, highly rated fragrances.

Leading Fragrance Influencer @JeremyFragrance

The roots of IBM’s research build on earlier computational creativity work used in the culinary domain, which was theory-driven and used flavor pairing models and psychological models of olfactory pleasantness to identify recipes that would taste and smell good. But, as opposed to a food recipe where amounts of ingredients can be vague or approximate, creating a fine fragrance is a science that requires precision as even the smallest percentage change in the amount of a material can make or break a new perfume. As a result, IBM took a data-driven approach, relying on information about hundreds of thousands of fragrance formulas, data on the fragrance families (e.g. fruity, oriental and floral), data on a fragrance’s raw materials and historical data that captured the success of previously designed perfumes and formulas, among other data points.

Using this mix of data, “Philyra” uses machine learning technologies to generate new combinations of fragrance formulations that fit the design objectives. For example: creating a unique perfume for Brazilian male millennials.

The components of the system specifically include algorithms that learn and predict:

• alternative raw material complements and substitutes that could be used in a formula
• the appropriate dosing for a raw material based on usage patterns
• the human response
• the novelty of the fragrance by comparing it to a large set of commercially available fragrances

When it comes to new fragrance formulation, novelty is a major driver, and the AI system learns a distance model to identify fragrances that are close in smell to existing fragrances. The larger the distance between a fragrance and its neighbors, the more novel the perfume is predicted to be since it occupies a white space in fragrance.

In responding to a brief from O Boticario, Symrise used the IBM Research system to design two new perfumes that will be rolled out next year. As part of the development process, the initial formulas suggested by the system were minimally tweaked by Symrise Perfumer David Apel to emphasize a certain fragrance note, and improve how long it lasts on the skin.

“Perfumery is an ancient art. It moves slowly through history and can seem unchanged for centuries. But there are moments of historic significance that are undeniable. The last great explosion of innovation began in the late 19th century with the introduction of synthetic fragrance raw materials into the perfumer’s pallet. That revolution is still the driving force behind most of the great perfumes of today. Artificial Intelligence is quite simply the next frontier of perfumery innovation and discovery. Watching it unfold, and helping to direct it, is a privilege and the most significant milestone in my career,” stated Apel.

Symrise plans to introduce “Philyra” to their perfumers around the globe.  

For more information on Symrise, visit: www.symrise.com

For more information on IBM Research, visit: http://www.research.ibm.com/