Identifying Forced Labor in Building Material Supply Chains
Project summary
This project uses interpretable machine learning methods to predict the risk of forced labor in mining operations for raw materials across various locations. The goal is to equip U.S. material specifiers and government officials with tools to mitigate the importation of materials linked to forced labor and unchecked environmental impacts. At the same time, the project will highlight high-risk areas, enabling NGOs and government agencies to conduct targeted investigations. Throughout this process we will develop an approach to address common challenges stemming from the scarcity and quality of existing forced labor data.
Driving questions
How can natural language processing (NLP) be used to augment existing forced labor data sets and create new ones?
What important predictors of forced labor risk can be identified using variable importance techniques and interpretable machine learning methods?
What is the connection between environmental impacts and forced labor risks?
Project team
Antonio Torres Skillicorn, CEE PhD student
Prof. Sarah Billington, CEE
Prof. Dan Iancu, Stanford Graduate School of Business
Our funders
Grace Farms Foundation (Design for Freedom)