Daniel Hsu - Computational Lower Bounds for Tensor PCA

preview_player
Показать описание
Invited talk at the Workshop on the Theory of Overparameterized Machine Learning (TOPML) 2022.
Speaker: Daniel Hsu (Columbia University)
Talk title: Computational Lower Bounds for Tensor PCA

  1. TOPML Workshop
Рекомендации по теме
Daniel Hsu - 0:55:06

Daniel Hsu - Computational Lower Bounds for Tensor PCA

Daniel Hsu: Transformers, 0:41:58

Daniel Hsu: Transformers, parallel computation and logarithmic depth

Contrastive learning, multi-view 0:11:03

Contrastive learning, multi-view redundancy, and linear models

Transformers, parallel computation, 0:57:21

Transformers, parallel computation, and logarithmic depth

Algorithms for Multi-Group 0:30:31

Algorithms for Multi-Group Learning

Introduction to Transformers 1:10:24

Introduction to Transformers

Tensor Decompositions for 0:55:49

Tensor Decompositions for Learning Latent Variable Models I

AI Institute 'Geometry 1:36:18

AI Institute 'Geometry of Deep Learning' 2019 [Workshop] Day 1 | Session 2

Graph Colouring Is 0:29:21

Graph Colouring Is Hard on Average for Polynomial Calculus and Nullstellensatz

Surprises in High-Dimensional 1:08:19

Surprises in High-Dimensional (Overparameterized) Linear Classification

Statistical Learning in 1:00:34

Statistical Learning in Operations: The Interplay between Online and Offline learning

Tensor Decompositions for 0:53:12

Tensor Decompositions for Learning Latent Variable Models II

Fit Without Fear: 0:55:13

Fit Without Fear: An Over-Fitting Perspective on Modern Deep and Shallow Learning

Strange Loop Language 0:37:39

Strange Loop Language Panel - Hickey, Sussman, Wirfs-Brock, Pamer, Alexandrescu, Ashkenas (2011)

The Platform Design 0:44:00

The Platform Design Problem

Unbiased Estimates for 0:35:47

Unbiased Estimates for Linear Regression via Volume Sampling

Transforming Health Care 0:48:57

Transforming Health Care with Technology: Model-Based Approaches

Day 1 - 1:21:47

Day 1 - M. Belkin: The puzzle of dimensionality and feature learning in modern Deep Learning and LLM

Daniel Kane - 0:45:58

Daniel Kane - UC San Diego - Ak Testing of Distributions

Arya Mazumdar - 0:56:06

Arya Mazumdar - Sample complexity of estimation in logistic regression - IPAM at UCLA

MagicDec: Breaking the 0:12:43

MagicDec: Breaking the Latency-Throughput Tradeoff for Long Context Generation with Spec

Dan Zuckerman - 0:33:01

Dan Zuckerman - Workshop on Free Energy, Kinetics, & MSM in Drug Design 2018

Representational Strengths and 0:58:01

Representational Strengths and Limitations of Transformers

HSU EGR 6163 0:22:08

HSU EGR 6163 Module 6 Lecture 1