COLT 2017: 30TH ANNUAL CONFERENCE ON LEARNING THEORY
PROGRAM

Days: Thursday, July 6th Friday, July 7th Saturday, July 8th Sunday, July 9th Monday, July 10th

Thursday, July 6th

View this program: with abstractssession overviewtalk overview

11:00-13:30Lunch Break
15:20-15:50Coffee Break
Friday, July 7th

View this program: with abstractssession overviewtalk overview

09:00-10:00 Session 5: Adaptivity and Human-centric Learning
09:00
Generalization for Adaptively-chosen Estimators via Stable Median ( abstract )
09:20
Learning Non-Discriminatory Predictors ( abstract )
09:40
The Price of Selection in Differential Privacy ( abstract )
09:50
Efficient PAC Learning from the Crowd ( abstract )
10:00-10:20Coffee Break
10:20-11:20 Session 6: Langevin Dynamics and Non-Convex Optimization
10:20
A Hitting Time Analysis of Stochastic Gradient Langevin Dynamics (Best Paper Award) ( abstract )
10:40
Non-Convex Learning via Stochastic Gradient Langevin Dynamics: A Nonasymptotic Analysis ( abstract )
10:50
Further and stronger analogy between sampling and optimization: Langevin Monte Carlo and gradient descent ( abstract )
11:00
Sampling from a log-concave distribution with compact support with proximal Langevin Monte Carlo ( abstract )
11:10
Fast Rates for Empirical Risk Minimization of Strict Saddle Problems ( abstract )
11:20-11:35Coffee Break
12:35-14:30Lunch Break (including Women in Machine Learning - Theory Lunch)
14:30-15:30 Session 8: Unsupervised Learning
14:30
Sample complexity of population recovery ( abstract )
14:50
Noisy Population Recovery from Unknown Noise ( abstract )
15:00
Learning Multivariate Log-concave Distributions ( abstract )
15:10
Ten Steps of EM Suffice for Mixtures of Two Gaussians ( abstract )
15:20
The Hidden Hubs Problem ( abstract )
15:30-16:00Coffee Break
16:00-17:00 Session 9: Bandits I
16:00
Sparse Stochastic Bandits ( abstract )
16:10
An Improved Parametrization and Analysis of the EXP3++ Algorithm for Stochastic and Adversarial Bandits ( abstract )
16:20
Corralling a Band of Bandit Algorithms ( abstract )
16:30
Lower Bounds on Regret for Noisy Gaussian Process Bandit Optimization ( abstract )
16:40
Towards Instance Optimal Bounds for Best Arm Identification ( abstract )
16:50
Bandits with Movement Costs and Adaptive Pricing ( abstract )
17:00-17:20Coffee Break
17:20-17:40 Session 10: Online Learning with Partial Feedback
17:20
Tight Bounds for Bandit Combinatorial Optimization ( abstract )
17:30
Online Nonparametric Learning, Chaining, and the Role of Partial Feedback ( abstract )
17:40-18:00 Session 11: Open Problems Session
17:40
Open Problem: First-Order Regret Bounds for Contextual Bandits ( abstract )
17:50
Open Problem: Meeting Times for Learning Random Automata ( abstract )
Saturday, July 8th

View this program: with abstractssession overviewtalk overview

09:00-10:00 Session 13: Robustness
09:00
Adaptivity to Noise Parameters in Nonparametric Active Learning ( abstract )
09:20
Computationally Efficient Robust Estimation of Sparse Functionals ( abstract )
09:30
Robust Proper Learning for Mixtures of Gaussians via Systems of Polynomial Inequalities ( abstract )
09:40
Ignoring Is a Bliss: Learning with Large Noise Through Reweighting-Minimization ( abstract )
09:50
Thresholding based Efficient Outlier Robust PCA ( abstract )
10:00-10:20Coffee Break
10:20-11:20 Session 14: Combinatorial Optimization in Learning
10:20
Solving SDPs for synchronization and MaxCut problems via the Grothendieck inequality ( abstract )
10:40
Learning-Theoretic Foundations of Algorithm Configuration for Combinatorial Partitioning Problems ( abstract )
10:50
Greed Is Good: Near-Optimal Submodular Maximization via Greedy Optimization ( abstract )
11:00
Submodular Optimization under Noise ( abstract )
11:10
Correspondence retrieval ( abstract )
11:20-11:35Coffee Break
11:35-12:35 Session 15: Online Learning
11:35
Online Learning Without Prior Information (Best Student Paper Award) ( abstract )
11:55
On Equivalence of Martingale Tail Bounds and Deterministic Regret Inequalities ( abstract )
12:15
Fast rates for online learning in Linearly Solvable Markov Decision Processes ( abstract )
12:25
ZIGZAG: A new approach to adaptive online learning ( abstract )
12:35-14:50Lunch Break
14:50-15:40 Session 17: PAC Learning
Chair:
14:50
Efficient Co-Training of Linear Separators under Weak Dependence ( abstract )
15:10
Effective Semisupervised Learning on Manifolds ( abstract )
15:20
Quadratic Upper Bound for Recursive Teaching Dimension of Finite VC Classes ( abstract )
15:30
Learning Disjunctions of Predicates ( abstract )
Sunday, July 9th

View this program: with abstractssession overviewtalk overview

09:00-10:00 Session 20: Complexity of Learning
09:00
A General Characterization of the Statistical Query Complexity ( abstract )
09:20
Mixing Implies Lower Bounds for Space Bounded Learning ( abstract )
09:40
On Learning versus Refutation ( abstract )
09:50
Inapproximability of VC Dimension and Littlestone's Dimension ( abstract )
10:00-10:20Coffee Break
10:20-11:20 Session 21: Property Testing and Elicitation
10:20
Memoryless Sequences for Differentiable Losses ( abstract )
10:40
Multi-Observation Elicitation ( abstract )
10:50
Testing Bayesian Networks ( abstract )
11:00
Square Hellinger Subadditivity for Bayesian Networks and its Applications to Identity Testing ( abstract )
11:10
Two-Sample Tests for Large Random Graphs using Network Statistics ( abstract )
11:20-11:35Coffee Break
12:35-14:30Lunch Break
14:30-15:30 Session 23: Stochastic Optimization
14:30
Empirical Risk Minimization for Stochastic Convex Optimization: $O(1/n)$- and $O(1/n^2)$-type of Risk Bounds ( abstract )
14:50
Stochastic Composite Least-Squares Regression with convergence rate O(1/n) ( abstract )
15:00
A Unified Analysis of Stochastic Optimization Methods Using Jump System Theory and Quadratic Constraints ( abstract )
15:10
Memory and Communication Efficient Distributed Stochastic Optimization with Minibatch Prox ( abstract )
15:20
The Sample Complexity of Optimizing a Convex Function ( abstract )
15:30-16:00Coffee Break
16:00-17:00 Session 24: Bandits II
Chair:
16:00
The Simulator: Understanding Adaptive Sampling in the Moderate-Confidence Regime ( abstract )
16:20
Learning with Limited Rounds of Adaptivity: Coin Tossing, Multi-Armed Bandits, and Ranking from Pairwise Comparisons ( abstract )
16:40
Nearly Optimal Sampling Algorithms for Combinatorial Pure Exploration ( abstract )
16:50
Thompson Sampling for the MNL-Bandit ( abstract )
17:00-17:20Coffee Break
Monday, July 10th

View this program: with abstractssession overviewtalk overview

09:00-10:00 Session 26: Neural Networks
09:00
On the Ability of Neural Nets to Express Distributions ( abstract )
09:20
Depth Separation for Neural Networks ( abstract )
09:30
Surprising properties of dropout in deep networks ( abstract )
09:40
Reliably Learning the ReLU in Polynomial Time ( abstract )
09:50
Nearly-tight VC-dimension bounds for neural networks ( abstract )
10:00-10:20Coffee Break
10:20-11:20 Session 27: Learning with Matrices and Tensors
Chair:
10:20
Exact tensor completion with sum-of-squares ( abstract )
10:40
Fast and robust tensor decomposition with applications to dictionary learning ( abstract )
10:50
Homotopy Analysis for Tensor PCA ( abstract )
11:00
Fundamental limits of symmetric low-rank matrix estimation ( abstract )
11:10
Matrix Completion from O(n) Samples in Linear Time ( abstract )
11:20-11:35Coffee Break
11:35-12:35 Session 28: Statistical Learning Theory
11:35
High-Dimensional Regression with Binary Coefficients. Estimating Squared Error and a Phase Transition. ( abstract )
11:55
Rates of estimation for determinantal point processes ( abstract )
12:05
Predicting with Distributions ( abstract )
12:15
A second-order look at stability and generalization ( abstract )
12:25
Optimal learning via local entropies and sample compression ( abstract )