Time-Series Deep Learning Model Predicts Flavor Development in Fengxiang-Style Taibai Baijiu

Written by

in

A study published in npj Science of Food used a temporal-fusion Transformer architecture to predict changes in flavors during the fermentation and aging of Fengxiang-style Taibai baijiu. The model showed that early microbial dynamics could predict subsequent changes in volatile compounds, while compound features were associated with the finished product’s later sensory properties. The study also validated associations involving 16 volatile compounds through sensory reconstruction experiments. However, these results represent temporal correlations and predictive relationships and cannot be directly interpreted as evidence of biological or chemical causal mechanisms.

Source:BIOENGINEER.ORG · Read Original

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *