High-Dimensional Robust Statistics with Ilias Diakonikolas - #351
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)
00:00
Navigating High-Dimensional Challenges
This chapter explores the difficulties of estimating mean values in high-dimensional spaces, particularly in the presence of outliers. It introduces the median as a more robust alternative and discusses an efficient algorithm for estimating means while mitigating the effects of outliers. The conversation also addresses the implications of noise in labeled data and the importance of different noise models on statistical learning outcomes.
Transcript
Play full episode
