Testing Dependency of Databases - Lecture
Offered By: Institute for Pure & Applied Mathematics (IPAM) via YouTube
Course Description
Overview
Explore a 45-minute conference talk on detecting dependency between random databases presented by Wasim Huleihel from Tel Aviv University at IPAM's EnCORE Workshop. Dive into the hypothesis testing problem where the null hypothesis assumes independent database generation, while the alternative posits dependency under a latent row permutation. Discover sharp thresholds for optimal testing error probability, examining how they relate to database dimensions and generative distributions. Gain insights into the phase transition from zero to one in error probability. Recorded on February 28, 2024, this presentation is part of the Computational vs Statistical Gaps in Learning and Optimization workshop at the Institute for Pure & Applied Mathematics (IPAM) at UCLA.
Syllabus
Wasim Huleihel - Testing Dependency of Databases - IPAM at UCLA
Taught by
Institute for Pure & Applied Mathematics (IPAM)
Related Courses
Introduction to Bayesian StatisticsDatabricks via Coursera Probability via Computation - Week 4
The Julia Programming Language via YouTube On the Unreasonable Effectiveness of Compressive Imaging - Ben Adcock, Simon Fraser University
Alan Turing Institute via YouTube The Mother of All Representer Theorems for Inverse Problems and Machine Learning - Michael Unser
Alan Turing Institute via YouTube Affine Spline Insights into Deep Learning - Richard Baraniuk, Rice University
Alan Turing Institute via YouTube