Optimization, Statistics and Uncertainty

source: Simons Institute      2017年11月28日
This workshop aims to explore the emerging connections between optimization, statistics and machine learning. It will discuss ties, both well-known and anticipated, between optimization and various approaches to modeling, reasoning about and coping with uncertainty. Some of the key topics that will be covered are: unsupervised machine learning, computational hardness in statistics, differential privacy and its applications in statistics and machine learning, online regret minimization, robust optimization and planted models.
For more information please visit: https://simons.berkeley.edu/workshops...
These presentations were supported in part by an award from the Simons Foundation.

1:07:53 Data Driven Optimization Models and Algorithms
32:12 Regularized Nonlinear Acceleration
31:44 Learning One-hidden-layer Neural Networks with Landscape Design
33:32 Algorithmic Stability for Interactive Data Analysis
27:00 Preventing Overfitting in Adaptive Data Analysis via Stability
30:41 The Convergence of Hamiltonian Monte Carlo
33:49 Distributional Robustness, Learning, and Empirical Likelihood
58:02 An Instability in Variational Methods for Learning Topic Models
32:33 A Few Connections Between Optimization and Probability
10 28:46 Implicit Regularization in Nonconvex Statistical Estimation
11 36:37 Low-rank Matrix Completion: Adaptive Sampling Can Help When, How?
12 1:01:40 Graph Powering
13 30:28 On Approximation Guarantees for Greedy Low Rank Optimization
14 29:53 Fast Rates for Bandit Optimization with Upper-Confidence Frank-Wolfe
15 1:01:19 Optimization's Untold Gift to Learning: Implicit Regularization
16 32:33 The Power and Limitations of Kernel Learning
17 29:27 Sampling Algorithms for Combinatorial Pure Exploration
18 59:07 Compositional Properties of Statistical Procedures: An Information-Theoretic View
19 28:02 Stability and Convergence Trade-Off of Iterative Optimization Algorithms
20 30:14 Accuracy First: Selecting a Differential Privacy Level for Accuracy-Constrained...

No comments: