Knowing When You Don't Know: Engineering AI Systems in an Uncertain World
December 2020 • Video
This presentation provides a view of new research about artificial intelligence (AI) system engineering and uncertainty.
Software Engineering Institute
SEI's Dr. Eric Heim discusses methods to quantify and reduce uncertainty in machine learning (ML) models. ML systems can make wrong predictions and give inaccurate estimates for the uncertainty of their predictions. On top of that, it is difficult to predict when their predictions will be wrong. This presentation at the 2020 SEI Research Review describes new techniques to quantify uncertainty, identify causes of uncertainty, and efficiently update ML models to reduce uncertainty in their predictions.