Sustainable biobased packaging via integrated robotics/machine learning workflow for improved postharvest preservation.

An integrated robotics–machine learning–DFT–LCA workflow predicts ~1 billion biobased packaging formulations, yielding antimicrobial, hydrostable, high-barrier films that extend cucumber and avocado shelf life.

• Why it matters: Petroleum-derived plastics increase environmental burdens and lack intrinsic antimicrobial protection, while biobased formulation optimization is slowed by huge component libraries and multiproperty tradeoffs.
• What they did: The team used data augmentation, ensemble modeling, and a hierarchical neural network to predict ~1 billion candidate formulations, then validated performance and assessed environmental footprint with life cycle assessment.
• The result: The resulting biobased packaging shows enhanced hydrostability, effective antimicrobial activity, and high barrier performance, significantly extending shelf life of cucumbers and avocados, and is released via a data-sharing platform.