The aims are Aims are broadly to improve the food safety and supply chain resilience of the poultry processing supply chain 1. Apply machine learning algorithms to search in process control data for predictors of very rare high-level Salmonella contamination events in finished products. 2. Build foundations of a federated learning approach for multiple processors to collectively, but appropriately privately, learn from and act on improved predictors of poultry safety. 3. Model supply chain response to rare contamination events that create supply reduction shocks to demonstrate how improved AI models can also improve poultry supply chain resiliency.