Differentially Quantized Gradient Methods

Channel:
United States Simons Institute for the Theory of Computing
Subscribers:
68,700
Published on ● Video Link: https://www.youtube.com/watch?v=vDH58m6qKAA


Duration: 33:55
357 views
12

Victoria Kostina (California Institute of Technology)
https://simons.berkeley.edu/talks/victoria-kostina-2023-05-25
Information-Theoretic Methods for Trustworthy Machine Learning



Other Videos By Simons Institute for the Theory of Computing


2023-06-06Random Restrictions and Fourier Concentration 2
2023-06-06Random Restrictions and Fourier Concentration 1
2023-06-06Introduction to Analysis of Boolean Functions 2
2023-06-05Introduction to Analysis of Boolean Functions 1
2023-06-03Session 2
2023-06-02Session 1
2023-05-26Learning Under Data Poisoning
2023-05-26Information-Theoretic Methods for Fair Risk Minimization
2023-05-26Overparameterized learning: an asymptotic toy model that shows interesting phenomena
2023-05-26Learning classification metrics from preference feedback
2023-05-26Differentially Quantized Gradient Methods
2023-05-26Data Valuation: A Perturbation-Aware Approach
2023-05-26Memorization in Machine Learning
2023-05-26Distribution estimation with user-level privacy and communication constraints
2023-05-25KEYNOTE: Information Constrained Optimal Transport
2023-05-25Privacy and Fairness in Collaborative AI
2023-05-25The Limits of Group Fairness and Predictive Multiplicity
2023-05-25Decision making with information-theoretic constraints
2023-05-25Information Thresholds in Structure Estimation
2023-05-25Fair, Explainable, and Lawful Machine Learning for High-Stakes Applications
2023-05-24KEYNOTE: Variational Formulations and Distributed Convex Optimization Methods for...


Tags:
Simons Institute
theoretical computer science
UC Berkeley
Computer Science
Theory of Computation
Theory of Computing
Information-Theoretic Methods for Trustworthy Machine Learning
Victoria Kostina