Humans perceive these flavors through biology sensors such as receptors and channels. If you and I can drink a chocolate-flavored milk, and we both can recognize that flavor as ‘chocolate’, then it is only natural to think there is some pattern-recognition mechanism behind flavors too.
At Noah’s Farm, we are defining a new field of study: AI Flavor Discovery, the science of representing flavor through numbers and creating it with generative AI.
AIFD
AI Flavor Discovery
/ˌeɪ aɪ ef ˈdiː/ · noun · new
Sense 1 · the measurement
Describing a flavor in numbers accurate enough that it can be compared, analysed and moved through computationally — a digital fingerprint for taste and aroma.
Two people agree a drink tastes of chocolate; AIFD asks what they are both measuring.
Sense 2 · the making
Using generative AI on that description to create flavors and materials the world has not tasted yet.
Under AIFD the target comes first, and the molecule is designed against it.
We are pioneers in making flavor computable. We represent flavor through numbers, connect biology with AI, and use generative models to create new formulations and molecules. We are building the science of what the world will taste next.
Noah’s Farm
People who know the industry
Xiaogen Yang, PhDScientific Advisor
Flavor is extraordinarily complex.
Progress demands a team that understands both the science and the craft. Noah’s Farm brings those strengths together around an ambition worth believing in: making flavor something we can deliberately design.
Per Jr. Greisen, PhDAdvisor · Computational protein design
Computation has already redrawn what is possible in protein design.
Flavor offers a compelling application for the capabilities behind AI-driven drug discovery: understanding molecular interactions with biology and designing toward a desired response. Bringing that scientific depth to flavor can establish an entirely new discipline: AI Flavor Discovery.
Build the future with us.
The field does not have its standards yet. Help write them.
Described accurately enough, a flavor is a single point. Reaching that point is the part that can go two ways: assemble it from materials that already exist, or design a molecule that does not.
Generate targeted formulations
Build toward your sensory target with the raw materials available to you. N-Block explores combinations and alternatives; N-OS brings that approach into your team’s R&D workflow.
Go beyond the ingredients that already exist. N-Palette designs new molecules toward your sensory and physicochemical needs, opening possibilities for co-development and licensing.
