Lawrence Livermore National Laboratory



Peer-Timo Bremer

Executive Summary

Through this basic science project, we seek to establish how and why certain best practices produce results in the new and rapidly expanding field of machine learning. This research supports the national missions that increasingly rely on machine learning, including terrorist threat reduction, by establishing explanations for complex models, designing network architectures to enable incremental learning, and developing abstractions of neural processes that can enable large-scale, self-supervised machine learning.