Knowing When You Don't Know: Quantifying and Reasoning about Uncertainty in Machine Learning Models

reportActive / Technical Report | Accesssion Number: AD1183582 | Open PDF

Abstract:

Our Work: Evaluating, Characterizing, Articulating, and Rectifying Uncertainty in ML models for the purpose of more informative and robust AI Systems.

Author(s):
Heim, Eric ; Kirchenbauer, John ; Helland, Jon ; Oaks, Jacob ; Singh, Aarti ; Lipton, Zachary
Author Organization(s):
CARNEGIE-MELLON UNIV PITTSBURGH PA
Funding Organization(s):
AIR FORCE LIFE CYCLE MANAGEMENT CENTER HANSCOM AFB MA, HANSCOM AFB , MA
Document Type:
Technical Report/Briefing Charts
Publication Date:
2022 Nov 14
Pagination:
29

Security Markings

DOCUMENT & CONTEXTUAL SUMMARY

Distribution Code:
A - Approved For Public Release
Distribution Statement: Public Release

RECORD

Collection: TRECMS
Identifying Numbers
Contract Number(s):
FA8702-15-D-0002
Subject Terms
Modernization Areas:
Autonomy
Communities of Interest:
No COI(s) Identified
Descriptor(s):
artificial intelligence, machine learning, detectors, software development, detection, classification, deep learning, information processing, learning, materials, national guard, probability, probability distributions, south carolina, training, uncertainty, universities
Keyword(s):
robust ai, modern object detection
Subject Categories:
Mathematical and Computer Sciences; Mathematical and Computer Sciences
Creation Date:
2022 Oct 31
Update Date:
2022 Dec 29