Articles
"I Wanted to Create Something Impossible for Human Hands alone" – Decorté's Challenge: Taking on "Cosmetic Formula Calculation" Using Quantum Computers
With a relentless pursuit of true luxury through advanced dermatological science and sensuous textures, Decorté has sparked a revolution in cosmetics development. For the first time in the world[1], a "quantum computer" was used to calculate product formulas for "pore care," a common skincare concern.
We interviewed researchers Kentaro Yoshikawa and Shun Obikane from KOSÉ Corporation to get a behind-the-scenes look at this innovation, which fuses science and human ingenuity.
PROFILE
Kentaro Yoshikawa
KOSÉ Laboratory, Makeup Product Research Lab
Joined the company in 2011, engaged in the development of skincare, haircare, and primarily cleansing products. Since 2019, has focused on quantum computer-related research at the Advanced Research Lab. Received the Best Oral Presentation Award at the 2024 Annual Congress for SCCJ for research on applying solubility parameters to powder dispersion technology.
PROFILE
Shun Obikane
KOSÉ Laboratory, Advanced Research Laboratory
Joined in 2020, now working as a data scientist after leveraging experience in AI-based image analysis research as a student. Tackles beauty challenges with mathematical approaches, including developing cosmetics using computers like quantum computers and developing new aging indicators through image analysis.
10 Seconds to Calculate 100 Billion Patterns[2]! Quantum Computer Technology Turning the Cosmetics Development Field Upside Down
What exactly is the difference between quantum computers and conventional computers?
YoshikawaIn a nutshell, the "calculating process" is completely different.
Conventional computers, while there's some innovation, basically have to process each possible combination one by one. In the development of cosmetic formulations, "combinations of ingredients and quantities" are the challenge—even a slight increase in the number of ingredients makes the number of combinations explode. If you tried brute-forcing all possibilities, some problems would take years to solve.
ObikaneQuantum computers, on the other hand, utilize the principles of "quantum mechanics" in their calculations. They can create a state where various combination possibilities are superimposed and calculated all at once.
There are several types of quantum computers, but one feature is their strength in combinatorial calculations. In formulation development, they make it possible to rapidly derive the optimal formula from a huge number of candidates.
How does this difference in calculation mechanism affect cosmetics development?
YoshikawaNormally, developing a cosmetic formula involves researchers deciding on ingredients and ratios in steps like "10% of ingredient A" or "20% of ingredient B," and then painstakingly checking each with a computer (conventional) and preparing physical samples.
ObikaneIn this project using quantum computers, we first defined the basic conditions for the formula and the desired ideal function.
All information about candidate ingredients is fully input, calculations run at high speed, and the "optimal combination of ingredients and their respective doses" is obtained.
Is the derived formula the single "best pattern"?
ObikaneWe can, of course, get multiple patterns depending on the settings, but we selected the very best among them.
In this project, within just 10 seconds, we calculated several dozen patterns—out of more than 100 billion possibilities—ranked by solubility. The researchers then evaluated them for keratin plug solubility and texture. Ultimately, the formula quantum computers identified as "best" was the one commercialized.
With conventional development, generating 100 billion options would be nearly impossible. This could transform how cosmetics are made.
An Industry-First: Quantifying How Easily Keratin Plugs Dissolve
ObikaneHowever, does this mean quantum computers and other computational devices can do everything? Not quite. To have computers perform calculations, you have to "quantify" the functionality of cosmetics.
But there are some elements that are difficult to quantify. That's why the technology Yoshikawa worked on—quantifying functions—is so crucial.
YoshikawaThis time, since we're talking about a "pore care" product, we quantified the "ease of dissolving keratin plugs." Keratin plugs are composed of proteins and lipids; we decided to target the "lipids."
YoshikawaTo quantify how easily the lipids dissolve, we independently developed what we call the "Keratin Plug Solubility Parameter."
Using three-dimensional coordinates, like latitude, longitude, and altitude on a map, the circle below represents the "solubility area" where keratin plugs dissolve easily.
How did you set up this "solubility area" for keratin plugs?
YoshikawaWe actually sampled keratin plugs from real people and quantitatively measured how much dissolved in different solvents—one by one.
Wait, actual human keratin plugs!?
YoshikawaYes, exactly. These are keratin plugs we collected manually from people who consented to take part. (laughs)
Thanks to this painstaking analysis, one of our biggest achievements was being able to quantify the "area where about 90% of people's keratin plugs dissolve."
ObikaneNow that the Keratin Plug Solubility Parameter can quantify things, the computer can select similar ingredients—in other words, "ingredients that dissolve easily."
Surprising Formulations Discovered by Quantum Computers—Beyond Human Imagination
YoshikawaTraditionally, developing a formula meant picking components similar to the lipids in keratin plugs and combining them. If the circle below represents the "solubility area," it would be ingredients like D or E.
So the closer a component is to the circle, the more similar it is, and thus the easier it dissolves?
YoshikawaExactly. On the other hand, people wouldn't pick components like A, B, or C since they're far from the circle, especially experienced researchers.
But if you combine A, B, and C in specific ratios, you can actually hit the dead center of the keratin plug solubility area.
Wow! It really does overlap with the very center—the area where keratin plugs dissolve best.
ObikaneComputers aren't held back by fixed ideas like human intuition or experience, so they can simply calculate to find "the best combinations and mixing ratios."
YoshikawaIn that sense, these are formulas humans would never have considered—completely beyond human intuition.
Fusing Human Sensibility & Technology—Unlocking New Potential in Cosmetics
Did you actually conduct experimental tests to verify the effectiveness, Yoshikawa?
YoshikawaYes, I did. I also tested it using artificial keratin plugs. During that period, Obikane would keep asking me, "So, how did it go?" (laughs)
ObikaneYou put in so much work prepping and setting everything up—I couldn't help but worry what would happen if the keratin plugs didn't dissolve after all that.
When did you feel you were onto something?
ObikaneThe moment the "artificial keratin plugs dissolved" with the formula created according to the Keratin Plug Solubility Parameter. Watching it break down right before my eyes, I thought, "Whoa, it really did dissolve!" (laughs)
Even when you're mathematically confident, it's mostly theoretical—so seeing an actual cosmetic product deliver on its function was both a shock and a thrill.
So that's how the Cosme Decorté AQ Pore Blackhead Dissolver was born—a product that truly fuses human expertise with science.
ObikaneSince quantum computers themselves are still a developing technology, we need to handle them with caution.
Also, when it comes to computation, people often imagine complete automation, but there are steps in the calculation process that require human adjustments.
YoshikawaEven before the calculation stage, 80% of the preparation is done by hand. This applies to various experiments as well, and especially when it comes to selecting fragrances and raw materials suitable for AQ's quality, these decisions are made by people.
We'd love to hear your outlook on future product development.
YoshikawaWe plan to continue utilizing quantum computers in the field of research and development.
I joined the company because I wanted to create cosmetics that could "change lives." By merging the power of people and science, I hope to create products that can be used by many.
ObikaneThrough this experience, I felt that it may be possible to control cosmetic functions mathematically.
Ultimately, I want to mathematically express "human aging" and connect that to the functions of cosmetics. There’s still a great deal to study and verify, but I’d be delighted if I can help establish new technologies.
[1] Calculation of ingredient selection and blending ratios using solubility parameters
[2] Selection of suitable ingredients and approximate blending ratios for cleansing
Text by Namiko Uno
How was the article?
Share
Related Articles
