Nestlé R&D is partnering with IBM Research to develop new tools that leverage the power of Artificial Intelligence and deep tech to bring breakthrough innovations to life. This research collaboration has led to the development of a generative AI tool that can identify novel high-barrier packaging materials. Packaging helps to protect food and beverages while also preventing food waste. Nestlé is continuously reducing the use of virgin plastic, pioneering the use alternative materials and novel technologies, and moving to recyclable mono-material and paper-based solutions. Identifying new packaging materials that meet each product's functional needs while ensuring food safety and quality is often meticulous. In some cases, this requires years of research. Nestlé and IBM scientists leveraged AI-based processing techniques to construct a knowledge base of known materials from public and proprietary documents. Subsequently, the team fine-tuned a fit-for-purpose chemical language model on this curated corpus, enabling it to learn the representation of the molecular structures. Using that knowledge, the teams leveraged the recently developed regression transformer by IBM Research to learn the correlation between key structural molecular features and the resulting physical-chemical properties. The resulting model can now propose entirely new high-barrier packaging materials that shield sensitive products from moisture, temperature swings and oxygen. | 
