Unbiased Estimates for Linear Regression via Volume Sampling

Published on ● Video Link: https://www.youtube.com/watch?v=Xh-25Xhtmp4



Duration: 35:47
716 views
6


Michal Derezinski (UC Santa Cruz)
https://simons.berkeley.edu/talks/unbiased-estimates-linear-regression-volume-sampling
Randomized Numerical Linear Algebra and Applications




Other Videos By Simons Institute for the Theory of Computing


2018-10-15Lower Bounds for Elimination via Weak Regularity
2018-10-15The Hardest Halfspace
2018-10-03The AI Trinity: Data + Algorithms + Infrastructure
2018-09-27Subspace Clustering using Ensembles of K-Subspaces
2018-09-27Spectrally Robust Graph Isomorphism
2018-09-27Fast Algorithms for Multivariate ODEs arising in Sampling and Learning
2018-09-27Matrix Martingales in Randomized Numerical Linear Algebra
2018-09-27A Fast Algorithm for Computing the Maximum Weight Base in a Linear Matroid
2018-09-27Matrix-free Construction of HSS Representations Using Adaptive Randomized Sampling
2018-09-26Smoothed Analysis of Low Rank Solutions to Semidefinite Programs via Burer Monteiro Factorization
2018-09-26Unbiased Estimates for Linear Regression via Volume Sampling
2018-09-26Adventures with Randomized Algebra for Extreme-Scale Signal Processing
2018-09-26Large Scale Stochastic Training of Neural Networks
2018-09-26Randomized Constrained Matrix Decompositions
2018-09-26Hessian Matrix Inversion in 10^10 Dimensions with Parametric Bootstraps
2018-09-26Ridge Regression and Deterministic Ridge Leverage Score Sampling
2018-09-25Scalable Algorithmic Primitives for Data Science
2018-09-25Error Estimation for Randomized Numerical Linear Algebra: Bootstrap Methods
2018-09-25Applying Randomized Methods for Zonotope Enumeration and...
2018-09-25Randomized Algorithms for Computing Full Matrix Factorizations
2018-09-25Convergence Theory for Iterative Eigensolvers



Tags:
Randomized Numerical Linear Algebra and Applications
Michal Derezinski
Simons Institute
Theory of Computing
Theory of Computation
Theoretical Computer Science
Computer Science
UC Berkeley