Training (Overparametrized) Neural Networks in Near-Linear Time

Training (Overparametrized) Neural Networks in Near-Linear Time

Channel:
United States Simons Institute for the Theory of Computing
Subscribers:
68,700
Published on ● Video Link: https://www.youtube.com/watch?v=tqzGuGpit_w


Duration: 18:55
320 views
3

12th Innovations in Theoretical Computer Science Conference (ITCS 2021)
http://itcs-conf.org/

Training (Overparametrized) Neural Networks in Near-Linear Time

Binghui Peng (Columbia University)
Zhao Song (Princeton)
Jan van den Brand (KTH Royal Institute of Technology)
Omri Weinstein (Columbia University)



Other Videos By Simons Institute for the Theory of Computing


2021-01-15Block Rigidity: Strong Multiplayer Parallel Repetition implies Super-Linear Lower Bounds for Turing
2021-01-15Lower Bounds for Off-Chain Protocols: Exploring the Limits of Plasma
2021-01-15Shrinkage of Decision Lists and DNF Formulas
2021-01-15Quantitative Correlation Inequalities via Semigroup Interpolation
2021-01-15Differentially Oblivious Turing Machines
2021-01-15Unknown I.I.D. Prophets: Better Bounds, Streaming Algorithms, and a New Impossibility
2021-01-15Online Paging with a Vanishing Regret
2021-01-15Towards local testability for quantum coding
2021-01-15Complete Problems for Multi-Pseudodeterministic Computations
2021-01-15A New Connection Between Node and Edge Depth Robust Graphs
2021-01-15Training (Overparametrized) Neural Networks in Near-Linear Time
2021-01-15Communication memento: Memoryless communication complexity
2021-01-15Relative Lipschitzness in Extragradient Methods and a Direct Recipe for Acceleration
2021-01-15An O(n) time algorithm for finding Hamilton cycles with high probability
2021-01-15Computational Complexity of the Hylland-Zeckhauser Scheme for One-Sided Matching Markets
2021-01-15Sensitivity Analysis of the Maximum Matching Problem
2021-01-15On Distributed Differential Privacy and Counting Distinct Elements
2021-01-15A Generalized Matching Reconfiguration Problem
2021-01-15Agnostic learning with unknown utilities
2021-01-15Quantum versus Randomized Communication Complexity, with Efficient Players
2021-01-15No quantum speedup over gradient descent for non-smooth convex optimization