AISTATS 2022 Schedule

Filter Events

MON 28 MAR

9 a.m.

Community Activities and Mentorship:

Affinity Groups Speed Networking Social

(ends 10:30 AM)

10:30 a.m.

Oral 1: Learning theory / General ML [10:30-11:30]

Orals 10:30-11:30

[10:30] Denoising and change point localisation in piecewise-constant high-dimensional regression coefficients

[10:45] Noise Regularizes Over-parameterized Rank One Matrix Recovery, Provably

[11:00] Survival regression with proper scoring rules and monotonic neural networks

[11:15] Multivariate Quantile Function Forecaster

(ends 11:30 AM)

11:30 a.m.

Oral 2: Bayesian methods / Sampling methods [11:30-12:30]

Orals 11:30-12:30

[11:30] Differentiable Bayesian inference of SDE parameters using a pathwise series expansion of Brownian motion

[11:45] Nonparametric Relational Models with Superrectangulation

[12:00] Robust Bayesian Inference for Simulator-based Models via the MMD Posterior Bootstrap

[12:15] Unifying Importance Based Regularisation Methods for Continual Learning

(ends 12:30 PM)

1:30 p.m.

Poster Session 1 [1:30-3:00]

Posters 1:30-3:00

On the Interplay between Information Loss and Operation Loss in Representations for Classification

On a Connection Between Fast and Sparse Oblivious Subspace Embeddings

Regret Bounds for Expected Improvement Algorithms in Gaussian Process Bandit Optimization

Adaptively Partitioning Max-Affine Estimators for Convex Regression

Nonparametric Relational Models with Superrectangulation

An Alternate Policy Gradient Estimator for Softmax Policies

Can we Generalize and Distribute Private Representation Learning?

Aligned Multi-Task Gaussian Process

Infinitely Deep Bayesian Neural Networks with Stochastic Differential Equations

Modeling Conditional Dependencies in Multiagent Trajectories

Conditionally Gaussian PAC-Bayes

Non-stationary Online Learning with Memory and Non-stochastic Control

Optimal partition recovery in general graphs

The Curse Revisited: When are Distances Informative for the Ground Truth in Noisy High-Dimensional Data?

Robustness and Reliability When Training With Noisy Labels

Permutation Equivariant Layers for Higher Order Interactions

On the Consistency of Max-Margin Losses

Neural Enhanced Dynamic Message Passing

Robust Probabilistic Time Series Forecasting

Triple-Q: A Model-Free Algorithm for Constrained Reinforcement Learning with Sublinear Regret and Zero Constraint Violation

MLDemon:Deployment Monitoring for Machine Learning Systems

Primal-Dual Stochastic Mirror Descent for MDPs

SAN: Stochastic Average Newton Algorithm for Minimizing Finite Sums

DEANN: Speeding up Kernel-Density Estimation using Approximate Nearest Neighbor Search

Tight bounds for minimum $\ell_1$-norm interpolation of noisy data

Crowdsourcing Regression: A Spectral Approach

Efficient interventional distribution learning in the PAC framework

Improved Algorithms for Misspecified Linear Markov Decision Processes

p-Generalized Probit Regression and Scalable Maximum Likelihood Estimation via Sketching and Coresets

Provable Continual Learning via Sketched Jacobian Approximations

Discovering Inductive Bias with Gibbs Priors: A Diagnostic Tool for Approximate Bayesian Inference

Obtaining Causal Information by Merging Datasets with MAXENT

SHAFF: Fast and consistent SHApley eFfect estimates via random Forests

Probing GNN Explainers: A Rigorous Theoretical and Empirical Analysis of GNN Explanation Methods

Federated Functional Gradient Boosting

Sampling in Dirichlet Process Mixture Models for Clustering Streaming Data

Robust Deep Learning from Crowds with Belief Propagation

Practical Schemes for Finding Near-Stationary Points of Convex Finite-Sums

Robust Bayesian Inference for Simulator-based Models via the MMD Posterior Bootstrap

Statistical Depth Functions for Ranking Distributions: Definitions, Statistical Learning and Applications

Last Layer Marginal Likelihood for Invariance Learning

ExactBoost: Directly Boosting the Margin in Combinatorial and Non-decomposable Metrics

AdaBlock: SGD with Practical Block Diagonal Matrix Adaptation for Deep Learning

Multi-armed Bandit Algorithm against Strategic Replication

Point Cloud Generation with Continuous Conditioning

PAC Top-$k$ Identification under SST in Limited Rounds

CATVI: Conditional and Adaptively Truncated Variational Inference for Hierarchical Bayesian Nonparametric Models

Rejection sampling from shape-constrained distributions in sublinear time

Fast and accurate optimization on the orthogonal manifold without retraction

Testing Granger Non-Causality in Panels with Cross-Sectional Dependencies

Standardisation-function Kernel Stein Discrepancy: A Unifying View on Kernel Stein Discrepancy Tests for Goodness-of-fit

On the Convergence Rate of Off-Policy Policy Optimization Methods with Density-Ratio Correction

Momentum Accelerates the Convergence of Stochastic AUPRC Maximization

Safe Active Learning for Multi-Output Gaussian Processes

Randomized Stochastic Gradient Descent Ascent

Semi-Implicit Hybrid Gradient Methods with Application to Adversarial Robustness

Probabilistic Numerical Method of Lines for Time-Dependent Partial Differential Equations

REPID: Regional Effect Plots with implicit Interaction Detection

Uncertainty Quantification for Bayesian Optimization

Noise Regularizes Over-parameterized Rank One Matrix Recovery, Provably

Optimal estimation of Gaussian DAG models

Weighted Gaussian Process Bandits for Non-stationary Environments

Non-separable Spatio-temporal Graph Kernels via SPDEs

Modelling Non-Smooth Signals with Complex Spectral Structure

LIMESegment: Meaningful, Realistic Time Series Explanations

Two-way Sparse Network Inference for Count Data

Gap-Dependent Bounds for Two-Player Markov Games

Regret, stability & fairness in matching markets with bandit learners

Random Effect Bandits

Grassmann Stein Variational Gradient Descent

Statistical and computational thresholds for the planted k-densest sub-hypergraph problem

Double Control Variates for Gradient Estimation in Discrete Latent Variable Models

Resampling Base Distributions of Normalizing Flows

Information-Theoretic Analysis of Epistemic Uncertainty in Bayesian Meta-learning

An Unsupervised Hunt for Gravitational Lenses

Sinkformers: Transformers with Doubly Stochastic Attention

Proximal Optimal Transport Modeling of Population Dynamics

Using time-series privileged information for provably efficient learning of prediction models

Off-Policy Risk Assessment for Markov Decision Processes

Multivariate Quantile Function Forecaster

Common Failure Modes of Subcluster-based Sampling in Dirichlet Process Gaussian Mixture Models - and a Deep-learning Solution

On Learning Mixture Models with Sparse Parameters

Pairwise Supervision Can Provably Elicit a Decision Boundary

Reward-Free Policy Space Compression for Reinforcement Learning

On Uncertainty Estimation by Tree-based Surrogate Models in Sequential Model-based Optimization

Survival regression with proper scoring rules and monotonic neural networks

State Dependent Performative Prediction with Stochastic Approximation

Multiple Importance Sampling ELBO and Deep Ensembles of Variational Approximations

Spectral Pruning for Recurrent Neural Networks

An Information-theoretical Approach to Semi-supervised Learning under Covariate-shift

Generative Models as Distributions of Functions

Learning Sparse Fixed-Structure Gaussian Bayesian Networks

VFDS: Variational Foresight Dynamic Selection in Bayesian Neural Networks for Efficient Human Activity Recognition

Reconstructing Test Labels from Noisy Loss Functions

(ends 3:00 PM)

3 p.m.

Oral 3: Causality / Trustworthy ML [3:00-4:00]

Orals 3:00-4:00

[3:00] Almost Optimal Universal Lower Bound for Learning Causal DAGs with Atomic Interventions

[3:15] Variance Minimization in the Wasserstein Space for Invariant Causal Prediction

[3:30] On the Assumptions of Synthetic Control Methods

[3:45] Differentially Private Densest Subgraph

(ends 4:00 PM)

4 p.m.

Community Activities and Mentorship:

MeMentor Mentoring

(ends 5:00 PM)

Oral 4: Bandits / Reinforcement learning [4:00-5:00]

Orals 4:00-5:00

[4:00] Optimal Rates of (Locally) Differentially Private Heavy-tailed Multi-Armed Bandits

[4:15] Nonstochastic Bandits and Experts with Arm-Dependent Delays

[4:30] Towards Agnostic Feature-based Dynamic Pricing: Linear Policies vs Linear Valuation with Unknown Noise

[4:45] Towards an Understanding of Default Policies in Multitask Policy Optimization

(ends 5:00 PM)

6 p.m.

Remarks:

Opening Remarks

(ends 6:15 PM)

6:15 p.m.

Invited Talk:

New Methods for Taming Endogeneity Bias in Observational Studies: Proxies, Bespoke Instruments and Invalid Instruments

Eric Tchetgen Tchetgen

(ends 7:15 PM)

7:15 p.m.

Poster Session 2 [7:15-8:45]

Posters 7:15-8:45

Sobolev Norm Learning Rates for Conditional Mean Embeddings

Reinforcement Learning with Fast Stabilization in Linear Dynamical Systems

Variational Marginal Particle Filters

Safe Optimal Design with Applications in Off-Policy Learning

Robust Training in High Dimensions via Block Coordinate Geometric Median Descent

Nuances in Margin Conditions Determine Gains in Active Learning

Towards Statistical and Computational Complexities of Polyak Step Size Gradient Descent

Is Bayesian Model-Agnostic Meta Learning Better than Model-Agnostic Meta Learning, Provably?

Spectral Robustness for Correlation Clustering Reconstruction in Semi-Adversarial Models

Sample Complexity of Robust Reinforcement Learning with a Generative Model

Two-Sample Test with Kernel Projected Wasserstein Distance

How and When Random Feedback Works: A Case Study of Low-Rank Matrix Factorization

Stochastic Extragradient: General Analysis and Improved Rates

Towards Agnostic Feature-based Dynamic Pricing: Linear Policies vs Linear Valuation with Unknown Noise

Model-free Policy Learning with Reward Gradients

Neural Contextual Bandits without Regret

Gap-Dependent Unsupervised Exploration for Reinforcement Learning

On perfectness in Gaussian graphical models

A New Notion of Individually Fair Clustering: $\alpha$-Equitable $k$-Center

Fast Sparse Classification for Generalized Linear and Additive Models

Parametric Bootstrap for Differentially Private Confidence Intervals

Triangular Flows for Generative Modeling: Statistical Consistency, Smoothness Classes, and Fast Rates

Learning Revenue-Maximizing Auctions With Differentiable Matching

Nearly Minimax Optimal Regret for Learning Infinite-horizon Average-reward MDPs with Linear Function Approximation

Learning Interpretable, Tree-Based Projection Mappings for Nonlinear Embeddings

Characterizing and Understanding the Generalization Error of Transfer Learning with Gibbs Algorithm

Optimal Accounting of Differential Privacy via Characteristic Function

Faster Unbalanced Optimal Transport: Translation invariant Sinkhorn and 1-D Frank-Wolfe

On the Convergence of Continuous Constrained Optimization for Structure Learning

Sample Complexity of Policy-Based Methods under Off-Policy Sampling and Linear Function Approximation

On Combining Bags to Better Learn from Label Proportions

PACm-Bayes: Narrowing the Empirical Risk Gap in the Misspecified Bayesian Regime

Finding Valid Adjustments under Non-ignorability with Minimal DAG Knowledge

Beyond the Policy Gradient Theorem for Efficient Policy Updates in Actor-Critic Algorithms

Independent Natural Policy Gradient always converges in Markov Potential Games

Deep Non-crossing Quantiles through the Partial Derivative

Optimal Dynamic Regret in Proper Online Learning with Strongly Convex Losses and Beyond

On the Global Optimum Convergence of Momentum-based Policy Gradient

Local SGD Optimizes Overparameterized Neural Networks in Polynomial Time

Decoupling Local and Global Representations of Time Series

Learning Competitive Equilibria in Exchange Economies with Bandit Feedback

Near Instance Optimal Model Selection for Pure Exploration Linear Bandits

Uncertainty Quantification for Low-Rank Matrix Completion with Heterogeneous and Sub-Exponential Noise

A Predictive Approach to Bayesian Nonparametric Survival Analysis

Iterative Alignment Flows

Look-Ahead Acquisition Functions for Bernoulli Level Set Estimation

Towards Federated Bayesian Network Structure Learning with Continuous Optimization

Wide Mean-Field Bayesian Neural Networks Ignore the Data

Norm-Agnostic Linear Bandits

Sharp Bounds for Federated Averaging (Local SGD) and Continuous Perspective

Measuring the robustness of Gaussian processes to kernel choice

Sensing Cox Processes via Posterior Sampling and Positive Bases

Towards Return Parity in Markov Decision Processes

Deep Multi-Fidelity Active Learning of High-Dimensional Outputs

Finding Dynamics Preserving Adversarial Winning Tickets

Low-Pass Filtering SGD for Recovering Flat Optima in the Deep Learning Optimization Landscape

Private Sequential Hypothesis Testing for Statisticians: Privacy, Error Rates, and Sample Size

Towards Understanding Biased Client Selection in Federated Learning

Orthogonal Multi-Manifold Enriching of Directed Networks

On the Convergence of Stochastic Extragradient for Bilinear Games using Restarted Iteration Averaging

Self-training Converts Weak Learners to Strong Learners in Mixture Models

Nearly Optimal Algorithms for Level Set Estimation

The Importance of Future Information in Credit Card Fraud Detection

Relational Neural Markov Random Fields

Asynchronous Upper Confidence Bound Algorithms for Federated Linear Bandits

Finding Nearly Everything within Random Binary Networks

Near-optimal Policy Optimization Algorithms for Learning Adversarial Linear Mixture MDPs

Variance Minimization in the Wasserstein Space for Invariant Causal Prediction

Tuning-Free Generalized Hamiltonian Monte Carlo

Strategic ranking

Threading the Needle of On and Off-Manifold Value Functions for Shapley Explanations

Multiway Spherical Clustering via Degree-Corrected Tensor Block Models

Computing D-Stationary Points of $\rho$-Margin Loss SVM

Flexible Accuracy for Differential Privacy

Hierarchical Bayesian Bandits

On the Value of Prior in Online Learning to Rank

Nonstationary multi-output Gaussian processes via harmonizable spectral mixtures

Online Page Migration with ML Advice

Federated Learning with Buffered Asynchronous Aggregation

Transductive Robust Learning Guarantees

The Curse of Passive Data Collection in Batch Reinforcement Learning

Second-Order Sensitivity Analysis for Bilevel Optimization

Structured Multi-task Learning for Molecular Property Prediction

Privacy Amplification by Subsampling in Time Domain

Federated Myopic Community Detection with One-shot Communication

Learning Inconsistent Preferences with Gaussian Processes

Fast Distributionally Robust Learning with Variance-Reduced Min-Max Optimization

Thompson Sampling with a Mixture Prior

On Coresets for Fair Regression and Individually Fair Clustering

Laplacian Constrained Precision Matrix Estimation: Existence and High Dimensional Consistency

Vanishing Curvature in Randomly Initialized Deep ReLU Networks

Margin-distancing for safe model explanation

Solving Marginal MAP Exactly by Probabilistic Circuit Transformations

Provable Lifelong Learning of Representations

Fast Fourier Transform Reductions for Bayesian Network Inference

Pareto Optimal Model Selection in Linear Bandits

Performative Prediction in a Stateful World

Mitigating Bias in Calibration Error Estimation

Fair Disaster Containment via Graph-Cut Problems

GalilAI: Out-of-Task Distribution Detection using Causal Active Experimentation for Safe Transfer RL

Preference Exploration for Efficient Bayesian Optimization with Multiple Outcomes

Policy Learning and Evaluation with Randomized Quasi-Monte Carlo

Online Competitive Influence Maximization

Predicting the utility of search spaces for black-box optimization: a simple, budget-aware approach

Variational Gaussian Processes: A Functional Analysis View

Moment Matching Deep Contrastive Latent Variable Models

Solving Multi-Arm Bandit Using a Few Bits of Communication

Amortized Rejection Sampling in Universal Probabilistic Programming

(ends 8:45 PM)

TUE 29 MAR

10 a.m.

Poster Session 3 [10:00-11:30]

Posters 10:00-11:30

Distributionally Robust Structure Learning for Discrete Pairwise Markov Networks

k-Pareto Optimality-Based Sorting with Maximization of Choice

Lagrangian manifold Monte Carlo on Monge patches

Extragradient Method: O(1/K) Last-Iterate Convergence for Monotone Variational Inequalities and Connections With Cocoercivity

Almost Optimal Universal Lower Bound for Learning Causal DAGs with Atomic Interventions

Lifted Primal-Dual Method for Bilinearly Coupled Smooth Minimax Optimization

Efficient Algorithms for Extreme Bandits

Ada-BKB: Scalable Gaussian Process Optimization on Continuous Domains by Adaptive Discretization

A Cramér Distance perspective on Quantile Regression based Distributional Reinforcement Learning

Optimal Rates of (Locally) Differentially Private Heavy-tailed Multi-Armed Bandits

Efficient and passive learning of networked dynamical systems driven by non-white exogenous inputs

Diversity and Generalization in Neural Network Ensembles

Are All Linear Regions Created Equal?

Particle-based Adversarial Local Distribution Regularization

Improved analysis of randomized SVD for top-eigenvector approximation

Deep Layer-wise Networks Have Closed-Form Weights

Sobolev Transport: A Scalable Metric for Probability Measures with Graph Metrics

Asymptotically Optimal Locally Private Heavy Hitters via Parameterized Sketches

PAC Learning of Quantum Measurement Classes : Sample Complexity Bounds and Universal Consistency

Dropout as a Regularizer of Interaction Effects

Forward Looking Best-Response Multiplicative Weights Update Methods for Bilinear Zero-sum Games

Basis Matters: Better Communication-Efficient Second Order Methods for Federated Learning

Identity Testing of Reversible Markov Chains

Predictive variational Bayesian inference as risk-seeking optimization

A Class of Geometric Structures in Transfer Learning: Minimax Bounds and Optimality

Minimal Expected Regret in Linear Quadratic Control

Corruption-robust Offline Reinforcement Learning

Learning Tensor Representations for Meta-Learning

Feature Collapsing for Gaussian Process Variable Ranking

On Multimarginal Partial Optimal Transport: Equivalent Forms and Computational Complexity

Certifiably Robust Variational Autoencoders

An Optimal Algorithm for Strongly Convex Minimization under Affine Constraints

Stateful Offline Contextual Policy Evaluation and Learning

Improving Attribution Methods by Learning Submodular Functions

A Witness Two-Sample Test

Nearly Tight Convergence Bounds for Semi-discrete Entropic Optimal Transport

Top K Ranking for Multi-Armed Bandit with Noisy Evaluations

Quadric Hypersurface Intersection for Manifold Learning in Feature Space

Spectral risk-based learning using unbounded losses

Learning to Plan Variable Length Sequences of Actions with a Cascading Bandit Click Model of User Feedback

Complex Momentum for Optimization in Games

CF-GNNExplainer: Counterfactual Explanations for Graph Neural Networks

Maillard Sampling: Boltzmann Exploration Done Optimally

Generalised GPLVM with Stochastic Variational Inference

Data Appraisal Without Data Sharing

Pick-and-Mix Information Operators for Probabilistic ODE Solvers

Differential privacy for symmetric log-concave mechanisms

Tile Networks: Learning Optimal Geometric Layout for Whole-page Recommendation

A Non-asymptotic Approach to Best-Arm Identification for Gaussian Bandits

MT3: Meta Test-Time Training for Self-Supervised Test-Time Adaption

Dimensionality Reduction and Prioritized Exploration for Policy Search

Mean Nyström Embeddings for Adaptive Compressive Learning

Predicting the impact of treatments over time with uncertainty aware neural differential equations.

Causal Effect Identification with Context-specific Independence Relations of Control Variables

Bayesian Classifier Fusion with an Explicit Model of Correlation

Bayesian Inference and Partial Identification in Multi-Treatment Causal Inference with Unobserved Confounding

Mode estimation on matrix manifolds: Convergence and robustness

Robust Stochastic Linear Contextual Bandits Under Adversarial Attacks

Transfer Learning with Gaussian Processes for Bayesian Optimization

Duel-based Deep Learning system for solving IQ tests

Warping Layer: Representation Learning for Label Structures in Weakly Supervised Learning

Encrypted Linear Contextual Bandit

New Coresets for Projective Clustering and Applications

Analysis of a Target-Based Actor-Critic Algorithm with Linear Function Approximation

Calibration Error for Heterogeneous Treatment Effects

LocoProp: Enhancing BackProp via Local Loss Optimization

Sequential Multivariate Change Detection with Calibrated and Memoryless False Detection Rates

Pairwise Fairness for Ordinal Regression

Best Arm Identification with Safety Constraints

Exact Community Recovery over Signed Graphs

Label differential privacy via clustering

Amortised Likelihood-free Inference for Expensive Time-series Simulators with Signatured Ratio Estimation

Finite Sample Analysis of Mean-Volatility Actor-Critic for Risk-Averse Reinforcement Learning

Efficient computation of the the volume of a polytope in high-dimensions using Piecewise Deterministic Markov Processes

Beyond Data Samples: Aligning Differential Networks Estimation with Scientific Knowledge

A Last Switch Dependent Analysis of Satiation and Seasonality in Bandits

Bias-Variance Decompositions for Margin Losses

Investigating the Role of Negatives in Contrastive Representation Learning

Learning Quantile Functions for Temporal Point Processes with Recurrent Neural Splines

Being a Bit Frequentist Improves Bayesian Neural Networks

Can Pretext-Based Self-Supervised Learning Be Boosted by Downstream Data? A Theoretical Analysis

Privacy Amplification by Decentralization

Estimators of Entropy and Information via Inference in Probabilistic Models

ContextGen: Targeted Data Generation for Low Resource Domain Specific Text Classification

Metalearning Linear Bandits by Prior Update

Can Functional Transfer Methods Capture Simple Inductive Biases?

Lifted Division for Lifted Hugin Belief Propagation

Online Continual Adaptation with Active Self-Training

How to scale hyperparameters for quickshift image segmentation

Jointly Efficient and Optimal Algorithms for Logistic Bandits

Denoising and change point localisation in piecewise-constant high-dimensional regression coefficients

Co-Regularized Adversarial Learning for Multi-Domain Text Classification

(ends 11:30 AM)

11:30 a.m.

Oral 5: Kernels / Optimization / Deep learning [11:30-12:30]

Orals 11:30-12:30

[11:30] Adversarially Robust Kernel Smoothing

[11:45] A Single-Timescale Method for Stochastic Bilevel Optimization

[12:00] Lifted Primal-Dual Method for Bilinearly Coupled Smooth Minimax Optimization

[12:15] Generative Models as Distributions of Functions

(ends 12:30 PM)

12:30 p.m.

Oral 6: Learning theory / Sampling methods [12:30-1:30]

Orals 12:30-1:30

[12:30] Amortized Rejection Sampling in Universal Probabilistic Programming

[12:45] Adaptive Importance Sampling meets Mirror Descent : a Bias-variance Tradeoff

[1:00] Loss as the Inconsistency of a Probabilistic Dependency Graph: Choose Your Model, Not Your Loss Function

[1:15] On the Consistency of Max-Margin Losses

(ends 1:30 PM)

2:30 p.m.

Community Activities and Mentorship:

Affinity Group Supported Pathways to ML Research Panel and Social

(ends 4:00 PM)

4 p.m.

Invited Talk:

Machine Learning for Ocean and Climate Modeling: Advances & Challenges

Laure Zanna

(ends 5:00 PM)

5 p.m.

Oral 7: Bayesian methods / Deep learning [5:00-6:00]

Orals 5:00-6:00

[5:00] Many processors, little time: MCMC for partitions via optimal transport couplings

[5:15] Rapid Convergence of Informed Importance Tempering

[5:30] Projection Predictive Inference for Generalized Linear and Additive Multilevel Models

[5:45] Density Ratio Estimation via Infinitesimal Classification

(ends 6:00 PM)

6:45 p.m.

Award Ceremony:

Award Ceremony

(ends 7:00 PM)

7 p.m.

Test Of Time:

Approximate Inference in Additive Factorial HMMs with Application to Energy Disaggregation

(ends 8:00 PM)

8 p.m.

Community Activities and Mentorship:

WiML - CWS Social

(ends 9:00 PM)

WED 30 MAR

9 a.m.

Oral 8: Learning theory / Sampling methods [9:00-10:00]

Orals 9:00-10:00

[9:00] Sampling from Arbitrary Functions via PSD Models

[9:15] Orbital MCMC

[9:30] Hardness of Learning a Single Neuron with Adversarial Label Noise

[9:45] Data-splitting improves statistical performance in overparameterized regimes

(ends 10:00 AM)

10 a.m.

Oral 9: Reinforcement learning / Deep learning [10:00-11:00]

Orals 10:00-11:00

[10:00] Beta Shapley: a Unified and Noise-reduced Data Valuation Framework for Machine Learning

[10:15] Faster Rates, Adaptive Algorithms, and Finite-Time Bounds for Linear Composition Optimization and Gradient TD Learning

[10:30] A general class of surrogate functions for stable and efficient reinforcement learning

[10:45] A Complete Characterisation of ReLU-Invariant Distributions

(ends 11:00 AM)

11:30 a.m.

Invited Talk:

Using ML to Discover the Underlying Models of Medicine

Mihaela van der Schaar

(ends 12:30 PM)

12:30 p.m.

Poster Session 4 [12:30-2:00]

Posters 12:30-2:00

Differentiable Bayesian inference of SDE parameters using a pathwise series expansion of Brownian motion

Federated Reinforcement Learning with Environment Heterogeneity

Conditional Gradients for the Approximately Vanishing Ideal

Improved Approximation Algorithms for Individually Fair Clustering

Unifying Importance Based Regularisation Methods for Continual Learning

Counterfactual Explanation Trees: Transparent and Consistent Actionable Recourse with Decision Trees

Derivative-Based Neural Modelling of Cumulative Distribution Functions for Survival Analysis

Optimal transport with $f$-divergence regularization and generalized Sinkhorn algorithm

On Some Fast And Robust Classifiers For High Dimension, Low Sample Size Data

Faster Rates, Adaptive Algorithms, and Finite-Time Bounds for Linear Composition Optimization and Gradient TD Learning

Exploiting Correlation to Achieve Faster Learning Rates in Low-Rank Preference Bandits

Provably Efficient Policy Optimization for Two-Player Zero-Sum Markov Games

Differentially Private Federated Learning on Heterogeneous Data

Adaptive A/B Test on Networks with Cluster Structures

Fixed Support Tree-Sliced Wasserstein Barycenter

Fast Rank-1 NMF for Missing Data with KL Divergence

Acceleration in Distributed Optimization under Similarity

Sampling from Arbitrary Functions via PSD Models

Gaussian Process Bandit Optimization with Few Batches

Increasing the accuracy and resolution of precipitation forecasts using deep generative models

Node Feature Kernels Increase Graph Convolutional Network Robustness

Adversarially Robust Kernel Smoothing

Identifiable Energy-based Representations: An Application to Estimating Heterogeneous Causal Effects

Equivariant Deep Dynamical Model for Motion Prediction

Nonstochastic Bandits and Experts with Arm-Dependent Delays

Adaptation of the Independent Metropolis-Hastings Sampler with Normalizing Flow Proposals

On Facility Location Problem in the Local Differential Privacy Model

Bayesian Link Prediction with Deep Graph Convolutional Gaussian Processes

On the complexity of the optimal transport problem with graph-structured cost

GraphAdaMix: Enhancing Node Representations with Graph Adaptive Mixtures

FLIX: A Simple and Communication-Efficient Alternative to Local Methods in Federated Learning

Debiasing Samples from Online Learning Using Bootstrap

Adaptive Gaussian Processes on Graphs via Spectral Graph Wavelets

Data-splitting improves statistical performance in overparameterized regimes

Marginalized Operators for Off-policy Reinforcement Learning

Coresets for Data Discretization and Sine Wave Fitting

Convergent Working Set Algorithm for Lasso with Non-Convex Sparse Regularizers

Learning Personalized Item-to-Item Recommendation Metric via Implicit Feedback

A Globally Convergent Evolutionary Strategy for Stochastic Constrained Optimization with Applications to Reinforcement Learning

Efficient Online Bayesian Inference for Neural Bandits

Neural score matching for high-dimensional causal inference

Multi-class classification in nonparametric active learning

Online Learning for Unknown Partially Observable MDPs

On Global-view Based Defense via Adversarial Attack and Defense Risk Guaranteed Bounds

SparseFed: Mitigating Model Poisoning Attacks in Federated Learning with Sparsification

Optimal channel selection with discrete QCQP

On the Generalization of Representations in Reinforcement Learning

On Distributionally Robust Optimization and Data Rebalancing

Effective Nonlinear Feature Selection Method based on HSIC Lasso and with Variational Inference

Learning a Single Neuron for Non-monotonic Activation Functions

On PAC-Bayesian reconstruction guarantees for VAEs

Variational Continual Proxy-Anchor for Deep Metric Learning

Physics Informed Deep Kernel Learning

Approximate Function Evaluation via Multi-Armed Bandits

Learning Pareto-Efficient Decisions with Confidence

Near-Optimal Task Selection for Meta-Learning with Mutual Information and Online Variational Bayesian Unlearning

Factorization Approach for Low-complexity Matrix Completion Problems: Exponential Number of Spurious Solutions and Failure of Gradient Methods

Differentially Private Densest Subgraph

Embedded Ensembles: infinite width limit and operating regimes

Disentangling Whether from When in a Neural Mixture Cure Model for Failure Time Data

Harmless interpolation in regression and classification with structured features

On Structured Filtering-Clustering: Global Error Bound and Optimal First-Order Algorithms

Feature screening with kernel knockoffs

A Spectral Perspective of DNN Robustness to Label Noise

Learning Proposals for Practical Energy-Based Regression

Parallel MCMC Without Embarrassing Failures

Identification in Tree-shaped Linear Structural Causal Models

A Single-Timescale Method for Stochastic Bilevel Optimization

Kantorovich Mechanism for Pufferfish Privacy

Adaptive Importance Sampling meets Mirror Descent : a Bias-variance Tradeoff

Near-optimal Local Convergence of Alternating Gradient Descent-Ascent for Minimax Optimization

Dual-Level Adaptive Information Filtering for Interactive Image Segmentation

Adaptive Multi-Goal Exploration

Implicitly Regularized RL with Implicit Q-values

Accurate Shapley Values for explaining tree-based models

On Margins and Derandomisation in PAC-Bayes

Differentially Private Regression with Unbounded Covariates

A cautionary tale on fitting decision trees to data from additive models: generalization lower bounds

Leveraging Time Irreversibility with Order-Contrastive Pre-training

How to Learn when Data Gradually Reacts to Your Model

Super-Acceleration with Cyclical Step-sizes

Exploring Image Regions Not Well Encoded by an INN

Diversified Sampling for Batched Bayesian Optimization with Determinantal Point Processes

On Linear Model with Markov Signal Priors

k-experts - Online Policies and Fundamental Limits

Structured variational inference in Bayesian state-space models

A Complete Characterisation of ReLU-Invariant Distributions

Marginalising over Stationary Kernels with Bayesian Quadrature

Generalized Group Testing

A general sample complexity analysis of vanilla policy gradient

Chernoff Sampling for Active Testing and Extension to Active Regression

System-Agnostic Meta-Learning for MDP-based Dynamic Scheduling via Descriptive Policy

Masked Training of Neural Networks with Partial Gradients

Model-agnostic out-of-distribution detection using combined statistical tests

(ends 2:00 PM)

3 p.m.

Oral 10: Gaussian processes / Optimization / Online ML [3:00-4:00]

Orals 3:00-4:00

[3:00] Minimax Optimization: The Case of Convex-Submodular

[3:15] Doubly Mixed-Effects Gaussian Process Regression

[3:30] Fast and Scalable Spike and Slab Variable Selection in High-Dimensional Gaussian Processes

[3:45] Debiasing Samples from Online Learning Using Bootstrap

(ends 4:00 PM)

4 p.m.

Oral 11: Learning theory / Kernels [4:00-5:00]

Orals 4:00-5:00

[4:00] Entropy Regularized Optimal Transport Independence Criterion

[4:15] Two-Sample Test with Kernel Projected Wasserstein Distance

[4:30] Estimating Functionals of the Out-of-Sample Error Distribution in High-Dimensional Ridge Regression

[4:45] Heavy-tailed Streaming Statistical Estimation

(ends 5:00 PM)

5:30 p.m.

Poster Session 5 [5:30-7:00]

Posters 5:30-7:00

Unlabeled Data Help: Minimax Analysis and Adversarial Robustness

A Bayesian Approach for Stochastic Continuum-armed Bandit with Long-term Constraints

Beta Shapley: a Unified and Noise-reduced Data Valuation Framework for Machine Learning

Synthsonic: Fast, Probabilistic modeling and Synthesis of Tabular Data

Meta Learning MDPs with linear transition models

Differentially Private Histograms under Continual Observation: Streaming Selection into the Unknown

Many processors, little time: MCMC for partitions via optimal transport couplings

Optimizing Early Warning Classifiers to Control False Alarms via a Minimum Precision Constraint

Online Control of the False Discovery Rate under “Decision Deadlines”

Efficient Kernelized UCB for Contextual Bandits

Offline Policy Selection under Uncertainty

QLSD: Quantised Langevin Stochastic Dynamics for Bayesian Federated Learning

Orbital MCMC

Adaptive Sampling for Heterogeneous Rank Aggregation from Noisy Pairwise Comparisons

Spiked Covariance Estimation from Modulo-Reduced Measurements

On Convergence of Lookahead in Smooth Games

Rapid Convergence of Informed Importance Tempering

Convex Analysis of the Mean Field Langevin Dynamics

Density Ratio Estimation via Infinitesimal Classification

Sketch-and-lift: scalable subsampled semidefinite program for K-means clustering

Learning and Generalization in Overparameterized Normalizing Flows

Outcome Assumptions and Duality Theory for Balancing Weights

Convergence of Langevin Monte Carlo in Chi-Squared and Rényi Divergence

PAC Mode Estimation using PPR Martingale Confidence Sequences

Conditionally Tractable Density Estimation using Neural Networks

The Fast Kernel Transform

Controlling Epidemic Spread using Probabilistic Diffusion Models on Networks

Conditional Linear Regression for Heterogeneous Covariances

Learning from Multiple Noisy Partial Labelers

Firebolt: Weak Supervision Under Weaker Assumptions

Efficient Hyperparameter Tuning for Large Scale Kernel Ridge Regression

A prior-based approximate latent Riemannian metric

Expressivity of Neural Networks via Chaotic Itineraries beyond Sharkovsky's Theorem

Minimax Optimization: The Case of Convex-Submodular

Zeroth-Order Methods for Convex-Concave Min-max Problems: Applications to Decision-Dependent Risk Minimization

Deep Generative model with Hierarchical Latent Factors for Time Series Anomaly Detection

Asynchronous Distributed Optimization with Stochastic Delays

Fundamental limits for rank-one matrix estimation with groupwise heteroskedasticity

The role of optimization geometry in single neuron learning

Polynomial Time Reinforcement Learning in Factored State MDPs with Linear Value Functions

Optimal Compression of Locally Differentially Private Mechanisms

Distributed Sparse Multicategory Discriminant Analysis

Compressed Rule Ensemble Learning

Entropy Regularized Optimal Transport Independence Criterion

Confident Least Square Value Iteration with Local Access to a Simulator

Entrywise Recovery Guarantees for Sparse PCA via Sparsistent Algorithms

Learning in Stochastic Monotone Games with Decision-Dependent Data

Towards an Understanding of Default Policies in Multitask Policy Optimization

A general class of surrogate functions for stable and efficient reinforcement learning

Faster One-Sample Stochastic Conditional Gradient Method for Composite Convex Minimization

Sample-and-threshold differential privacy: Histograms and applications

On the Implicit Bias of Gradient Descent for Temporal Extrapolation

On the Oracle Complexity of Higher-Order Smooth Non-Convex Finite-Sum Optimization

Causally motivated shortcut removal using auxiliary labels

A View of Exact Inference in Graphs from the Degree-4 Sum-of-Squares Hierarchy

Communication-Compressed Adaptive Gradient Method for Distributed Nonconvex Optimization

A Bandit Model for Human-Machine Decision Making with Private Information and Opacity

A Unified View of SDP-based Neural Network Verification through Completely Positive Programming

Provable Adversarial Robustness for Fractional Lp Threat Models

Optimal Design of Stochastic DNA Synthesis Protocols based on Generative Sequence Models

An Online Learning Approach to Interpolation and Extrapolation in Domain Generalization

Faster Single-loop Algorithms for Minimax Optimization without Strong Concavity

Policy Learning for Optimal Individualized Dose Intervals

Doubly Mixed-Effects Gaussian Process Regression

Fast and Scalable Spike and Slab Variable Selection in High-Dimensional Gaussian Processes

Parameter-Free Online Linear Optimization with Side Information via Universal Coin Betting

Hardness of Learning a Single Neuron with Adversarial Label Noise

TD-GEN: Graph Generation Using Tree Decomposition

Recoverability Landscape of Tree Structured Markov Random Fields under Symmetric Noise

Common Information based Approximate State Representations in Multi-Agent Reinforcement Learning

On the equivalence of Oja's algorithm and GROUSE

Scaling and Scalability: Provable Nonconvex Low-Rank Tensor Completion

Hypergraph Simultaneous Generators

Variational Autoencoders: A Harmonic Perspective

Learning from an Exploring Demonstrator: Optimal Reward Estimation for Bandits

On the Assumptions of Synthetic Control Methods

The Tree Loss: Improving Generalization with Many Classes

Outlier-Robust Optimal Transport: Duality, Structure, and Statistical Analysis

A Bayesian Model for Online Activity Sample Sizes

Learning Quantile Functions without Quantile Crossing for Distribution-free Time Series Forecasting

Weak Separation in Mixture Models and Implications for Principal Stratification

Deep Neyman-Scott Processes

A Dual Approach to Constrained Markov Decision Processes with Entropy Regularization

Cross-Loss Influence Functions to Explain Deep Network Representations

Approximate Top-$m$ Arm Identification with Heterogeneous Reward Variances

A Contraction Theory Approach to Optimization Algorithms from Acceleration Flows

Contrasting the landscape of contrastive and non-contrastive learning

Loss as the Inconsistency of a Probabilistic Dependency Graph: Choose Your Model, Not Your Loss Function

A Manifold View of Adversarial Risk

Pulling back information geometry

Estimating Functionals of the Out-of-Sample Error Distribution in High-Dimensional Ridge Regression

An Information-Theoretic Justification for Model Pruning

Adaptive Private-K-Selection with Adaptive K and Application to Multi-label PATE

Cycle Consistent Probability Divergences Across Different Spaces

Projection Predictive Inference for Generalized Linear and Additive Multilevel Models

Heavy-tailed Streaming Statistical Estimation

Adversarial Tracking Control via Strongly Adaptive Online Learning with Memory

Minimax Kernel Machine Learning for a Class of Doubly Robust Functionals with Application to Proximal Causal Inference

Equivariance Discovery by Learned Parameter-Sharing

A Dimensionality Reduction Method for Finding Least Favorable Priors with a Focus on Bregman Divergence

A Random Matrix Perspective on Mixtures of Nonlinearities in High Dimensions

One-bit Submission for Locally Private Quasi-MLE: Its Asymptotic Normality and Limitation

Exploring Counterfactual Explanations Through the Lens of Adversarial Examples: A Theoretical and Empirical Analysis

Convergence of online k-means

(ends 7:00 PM)

7 p.m.

Community Activities and Mentorship:

Mentoring Retrospectives Workshop

(ends 9:00 PM)