CONFERENCE PROCEEDINGS
Advanced Search
Home > Proceedings > Volume PC12117 > Article
Presentation
30 May 2022 When to retrain? assessing production ML model performance using uncertainty, out-of-distribution detection, and concept drift detection
James Holt
Author Affiliations +
Proceedings Volume PC12117, Disruptive Technologies in Information Sciences VI; PC1211702 (2022) https://doi.org/10.1117/12.2618321
Event: SPIE Defense + Commercial Sensing, 2022, Orlando, Florida, United States
Abstract
Submitting this abstract for an invited talk. The Laboratory for Physical Sciences has recently been conducting research in ML model uncertainty and confidence, detecting out-of-distribution data, and detecting concept drift. As we deploy ML models into operations, we must be constantly assessing whether the models are still effective and performing as expected in the current data environment. This is relevant in all cases, but especially critical in cybersecurity applications, because the data, technology, actors and behaviors are all evolving so rapidly. This talk will review several algorithmic techniques developed to address this problem.
Conference Presentation
© (2022) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Citation Download Citation
James Holt "When to retrain? assessing production ML model performance using uncertainty, out-of-distribution detection, and concept drift detection", Proc. SPIE PC12117, Disruptive Technologies in Information Sciences VI, PC1211702 (30 May 2022); https://doi.org/10.1117/12.2618321
ACCESS THE FULL ARTICLE
ORGANIZATIONAL
Sign in with credentials provided by your organization.
INSTITUTIONAL
Select your institution to access the SPIE Digital Library.
SELECT YOUR INSTITUTION
PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.
PERSONAL SIGN IN
No SPIE Account? Create one
PURCHASE THIS CONTENT
SUBSCRIBE TO DIGITAL LIBRARY
50 downloads per 1-year subscription
Members: $195
Non-members: $335 ADD TO CART
25 downloads per 1 - year subscription
Members: $145
Non-members: $250 ADD TO CART
PURCHASE SINGLE ARTICLE
Includes PDF, HTML & Video, when available
Members: $17.00
Non-members: $21.00 ADD TO CART
PROCEEDINGS
 PRESENTATION
WATCH
PRESENTATION SAVE TO MY LIBRARY
GET CITATION
Advertisement
Advertisement
KEYWORDS
Data modeling
Performance modeling
Algorithm development
Physical sciences
RELATED CONTENT
Subscribe to Digital Library
Receive Erratum Email Alert
Back to Top