Sugar function manifold: physics-gated AI for multi-axis sugar reduction

Overview

The Sugar Function Manifold (SFM) is an innovative AI-powered solution designed to revolutionize sugar reduction in the food industry. By recognizing that sugar influences not just sweetness but also other critical functions like water activity, flavor, and stability, SFM addresses the problem on multiple axes. This approach ensures that sugar-reduced products, such as biscuits, cakes, and confectionery, retain their sensory qualities and shelf stability.

Technical specifications

The core technology of SFM involves a physics-gated AI optimization process. It maps ingredients according to seven key sugar functions: sweetness, water activity, glass transition, Maillard browning, viscoelasticity, flavor, and bulk. By utilizing thermodynamic principles like the Gordon-Taylor and Maillard kinetics as physics gates, it filters out infeasible combinations, reducing the search space by 65% before AI optimization. The multi-axis swarm optimizer then identifies clean-label ingredient combinations that achieve at least 30% sugar reduction without compromising on quality. SFM is scale-ready, leveraging ingredients such as Allulose, Reb M, and rare sugars.

Technology readiness level

The Sugar Function Manifold is currently at Technology Readiness Level 3. The methodology has been validated in complex electrochemical systems and is domain-agnostic, making it applicable to food formulation. A six-month research phase is planned to develop a fully deployable tool, utilizing existing lab data and resources.