Machine learning ensembles surpassed mechanistic baselines in predicting Salmonella inactivation

NTU Food Research Systems (NTU‑FRS) congratulates Prof William Chen on the publication of a new research article in Food Chemistry Advances, titled “Predictive Modeling of Salmonella Inactivation in a Hybrid Alternative Protein Matrix: A Comparative Evaluation of Mechanistic and Machine Learning Approaches”.

The study explores whether machine learning models can outperform classical models in predicting Salmonella inactivation in alternative protein matrices using oregano essential oil as a natural antimicrobial. The work identifies a hybrid modelling path – one that combines mechanistic interpretability with machine learning predictive power – highlighting a promising direction for AI-driven food safety monitoring and management.

This publication represents an early research outcome of the FRESH × AWS Joint Lab, established under the Future Ready Food Safety Hub (FRESH) @ NTU in partnership with Amazon Web Services (AWS) through the Cloud to Table initiative. The joint efforts are driving the development of ML‑powered predictive microbiology from concept to practice.

Youssef Ezzaky, Mariem Zanzan, Amanda Voo Ying Hui, Fouad Achemchem, Wei Ning Chen, “Predictive modeling of Salmonella inactivation in a hybrid alternative protein matrix: A comparative evaluation of mechanistic and machine learning approaches”, Food Chemistry Advances, Volume 10, 2026, 101261,

DOI: https://doi.org/10.1016/j.focha.2026.101261