Sensory Analysis with AI (SENSAI)

Important results from the project

The project delivered an PoC AI-modell that predicts sensory properties from chemical analysis data. The model showed strong predictive ability and identified important chemical contributors to flavor and intensity. A user-friendly interface solution was also developed. The objectives were met by demonstrating that AI can be used for sensory analysis and by developing a working prototype.

Expected long term effects

The project is expected to streamline sensory quality control and product development in packaging materials in the long term by using AI to predict taste and intensity based on chemical analyses. The results enable faster testing of new raw materials and processes, which accelerates innovation. New molecules and groups have been identified that better explain sensory properties, opening up a more data-driven approach and new research opportunities.

Approach and implementation

The approach included the development of a prototype in agile working methodology and testing of the model. The project developed as planned and the main goals were met, although some limitations were identified. The schedule was largely followed, with more time allocated to the data work than initially planned. The collaboration between the parties worked well and led to new networks and research initiatives.