Research Review 2020, Knowing When You Don't Know: Engineering AI Systems in an Uncertain World

Abstract

In order for the DoD to leverage recent advances in AI, modern Machine Learning techniques need to be able to quantify, reason about, and rectify uncertainty in their predictions. In this work, we will benchmark modern techniques that quantify uncertainty, and develop techniques to identify causes of uncertainty and efficiently update ML models to reduce uncertainty in their predictions.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 2020
Accession Number
AD1111265

Entities

People

  • Eric Heim

Organizations

  • Carnegie Mellon University

Tags

DTIC Thesaurus Topics

  • Artificial Intelligence
  • Best Practices
  • Data Sets
  • Department Of Defense
  • Distance Learning
  • Engineering
  • Governments
  • Guarantees
  • Learning
  • Machine Learning
  • Materials
  • Neural Networks
  • Software Development
  • Training
  • Uncertainty
  • Universities

Fields of Study

  • Computer science

Readers

  • Neural Network Machine Learning.
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - DoD AI Strategy
  • AI & ML - Neural Networks