061: Interpolation, Extrapolation and Linearisation (Prof. Yann LeCun, Dr. Randall Balestriero)
Machine Learning Street Talk (MLST)
00:00
The Geometry of Learning: Interpolation vs. Extrapolation
This chapter examines the critical concepts of interpolation and extrapolation in machine learning, investigating their definitions and implications in high-dimensional spaces. It highlights the challenges posed by new data points in relation to known data and introduces alternative geometric frameworks for understanding these relationships.
Play episode from 57:22
Transcript
