Assumption-Free Prediction Intervals for Black-Box Regression Algorithms - Aaditya Ramdas
Institute for Advanced Study via YouTube
Pass the PMP® Exam on Your First Try — Expert-Led Training
Free courses from frontend to fullstack and AI
Overview
Google, IBM & Meta Certificates – 40% Off
One plan covers every Professional Certificate on Coursera.
Unlock All Certificates
Explore assumption-free prediction intervals for black-box regression algorithms in this comprehensive seminar on Theoretical Machine Learning. Delve into the distinctions between prediction intervals and confidence intervals, and understand the importance of predictive influence. Examine various classes of methods, including split conformal prediction, with practical examples. Learn about Conformalized Quantile Regression (CQR) and full conformal prediction techniques. Discover the application of jackknife intervals through illustrative examples. Gain insights from speaker Aaditya Ramdas of Carnegie Mellon University as he presents cutting-edge research in this 1-hour 22-minute talk at the Institute for Advanced Study.
Syllabus
Introduction
What is prediction
Why predictive influence
Prediction interval vs confidence interval
Assumptionfree predictive influence
Classes of methods
Split conformal prediction
Split conformal example
CQR
Full conformal prediction
Jackknife interval
Example
Taught by
Institute for Advanced Study