One of the key uses of AI in supply chains is the prediction of supply chain risk. Our lab proposed the use of ERP datasets for predicting order delays, which was adopted by many companies. Some of our current projects include:

Privacy preserving collective risk prediction: Funded by the EPSRC and in collaboration with aerospace and maritime industries, we examine the role of federated machine learning for organisations to predict risk collectively. For example, companies that share suppliers may pool their historical knowledge to estimate the likelihood of delivery delays, without having to share underlying sensitive data. Read our latest research on this paradigm here.

Digital supply chain surveillance refers to the proactive monitoring and analysis of digital data that allows firms to extract information related to a supply chain network, without needing the explicit consent of firms involved in the supply chain. AI has made DSCS  to become easier and larger-scale, posing significant opportunities for automated detection of actors and dependencies involved in a supply chain, which in turn, can help firms to detect risky, unethical and  environmentally unsustainable practices. Amongst many of its uses, is the use of both internal and external data sources on the prediction of supply risk.

Ingredients in processed products may be “hidden”, which in turn might disguise their origin, making it difficult for an organisation ascertain the authenticity of a product and its ethical credentials. We are working on predicting hidden ingredients, based on graph neural network techniques.

Predicting product quality: One of the key challenges in supply chains is improving the accuracy of quality monitoring. We are working on generative learning approaches that can extract rich information from production line records, thereby eliminating the need for a large numbers of samples that were produced under normal operational conditions.