PROGRAM
Days: Wednesday, September 28th Thursday, September 29th Friday, September 30th
Wednesday, September 28th
View this program: with abstractssession overviewtalk overview
08:30-10:10 Session WeA1: Intersection of Learning, Optimization, and Control I
Organizer: Bin Hu
Location: Library
| 08:30 | Convex Parameterization of Stabilizing Controllers and its LMI-based Computation via Filtering (abstract) |
| 08:50 | Regularized Gradient Descent Ascent for Two-Player Zero-Sum Markov Games (abstract) |
| 09:10 | TaSIL: Taylor Series Imitation Learning (abstract) |
| 09:30 | Optimal Private Data Acquisition: Central and Local Differential Privacy (abstract) |
08:30-10:10 Session WeA2: High Dimensional Statistics and Learning I
Organizers: Bruce Hajek, Jiaming Xu, Yihong Wu
Chair:
Yuguo Chen (University of Illinois at Urbana-Champaign, United States)
Location: Solarium
| 08:30 | Leave-one-out Singular Subspace Perturbation Analysis for Spectral Clustering (abstract) |
| 08:50 | Near-optimal algorithms for Imitation Learning (abstract) |
| 09:10 | Selecting the Number of Communities in Count-Weighted Networks (abstract) |
08:30-10:10 Session WeA3: Learning I
Location: Butternut
| 08:30 | Active Learning for Individual Data via Minimal Stochastic Complexity (abstract) |
| 08:50 | Byzantine Resilience With Reputation Scores (abstract) |
| 09:10 | CoBAAF: Controlled Bayesian Air Aggregation Federated Learning from Heterogeneous Data (abstract) |
| 09:30 | Ensuring the Defense of Paths and Perimeters in Dynamic Defender-Attacker Blotto Games (dDAB) on Graphs (abstract) PRESENTER: Bryce Ferguson |
| 09:50 | Over-the-Air Federated Learning with Privacy Protection via Correlated Additive Perturbations (abstract) PRESENTER: Erik G. Larsson |
08:30-10:10 Session WeA4: Information Theory and Coding I
Location: Pine
| 08:30 | Permutation Codes for Correcting a Burst of at Most $t$ Deletions (abstract)   PRESENTER: Yuanyuan Tang |
| 08:50 | Bounds on Reversible, Complement, Reversible-Complement, Constant Weight Sum Codes (abstract) |
| 09:10 | Zero-error Coding for Computing with Encoder Side-information (abstract) |
| 09:30 | Extracting Unique Information Through Markov Relations (abstract) |
| 09:50 | Strategic Communication via Cascade Multiple-Description Network (abstract) |
10:30-12:30 Session WeB1: Intersection of Learning, Optimization, and Control II
Organizer: Bin Hu
Location: Library
| 10:30 | Momentum-Based Learning and Optimization Dynamics with Stochastic Restarting (abstract) |
| 10:50 | Stability and Safety Constrained Reinforcement Learning for Voltage Control (abstract) |
| 11:10 | Representation-assisted Reinforcement Learning under Distribution Shift (abstract) |
| 11:30 | Mixed Strategies with Finite Support in Continuous Spaces: A Case Study in Trajectory Games (abstract) |
10:30-12:30 Session WeB2: Sequential Methods I
Organizer: G. Fellouris and V. Veeravalli
Location: Solarium
| 10:30 | Quickest Detection of the Change of Community via Stochastic Block Models (abstract) |
| 10:50 | Spatio-temporal Multiple Change-point Detection in Sensor Networks (abstract) |
| 11:10 | Quickest Detection of a Threat to an Impending Disaster (abstract) |
| 11:30 | Online detection of cascading change-points using diffusion networks (abstract) |
| 11:50 | Window-Limited CUSUM for Sequential Change Detection (abstract) |
10:30-12:30 Session WeB3: Networks, Learning, and Algorithms I
Organizers: Bruce Hajek and R. Srikant
Chair:
Thinh Doan (Virginia Tech, United States)
Location: Butternut
| 10:30 | Decentralized Cooperative Reinforcement Learning with Hierarchical Information Structure (abstract) |
| 10:50 | Bandits with Dynamic Arm-acquisition Costs (abstract) |
| 11:10 | A Survey of Dynamic Watermarking Algorithms and Some Recent Developments (abstract) |
| 11:30 | Leveraging Spatial and Temporal Correlations in Distributed Learning (abstract) |
10:30-12:30 Session WeB4: Information Theory and Coding II
Location: Pine
| 10:30 | Uncertainty in Biometric Identification and Authentication Systems with Strong Secrecy (abstract) |
| 10:50 | Linear Coding for Gaussian Two-Way Channels (abstract) PRESENTER: Junghoon Kim |
| 11:10 | Finite-Blocklength Results for the A-channel: Applications to Unsourced Random Access and Group Testing (abstract) |
| 11:30 | On the Symmetries of the Deletion Channel (abstract) |
13:30-15:10 Session WeC1: Networks, Learning, and Algorithms II
Organizers: Bruce Hajek and R. Srikant
Chair:
Assaf Zeevi (Columbia University, United States)
Location: Library
| 13:30 | Beyond No Regret: Instance-Dependent PAC Reinforcement Learning (abstract) |
| 13:50 | Repeated Games, Optimal Channel Capture, and Open Problems for Slotted Multiple Access (abstract) |
| 14:10 | The Multisecretary Problem: Regret, Approximation and Learning (abstract) |
| 14:30 | Markovian Interference in Experiments (abstract) |
13:30-15:10 Session WeC2: Sequential Methods II
Organizers: G. Fellouris and V. Veeravalli
Location: Solarium
| 13:30 | Efficient SPRT-based Best Arm Identification in Stochastic Bandits (abstract) |
| 13:50 | Efficient sequential UCB-based Hungarian algorithm for assignment problems (abstract) |
| 14:10 | Data-Driven Sequential Change Detection in Privacy-Sensitive Networks (abstract) PRESENTER: Yasin Yilmaz |
| 14:30 | Decentralized Anomaly Detection via Deep Multi-Agent Reinforcement Learning (abstract) |
13:30-15:10 Session WeC3: Representation, Learning, and Inference
Location: Butternut
| 13:30 | Oversquashing in GNNs through the lens of information contraction and graph expansion (abstract) PRESENTER: Kedar Karhadkar |
| 13:50 | Classifying Incomplete Data with a Mixture of Subspace Experts (abstract) |
| 14:10 | Multivariate Feature Extraction (abstract) PRESENTER: Xiangxiang Xu |
| 14:30 | On Sample Complexity of Learning Shared Representations: The Asymptotic Regime (abstract) PRESENTER: Xiangxiang Xu |
| 14:50 | Distributed Consensus-based Kalman Filter under Limited Communication (abstract) |
13:30-15:10 Session WeC4: Learning II
Location: Pine
| 13:30 | Decentralized Online Influence Maximization (abstract) |
| 13:50 | Learning-based Optimal Admission Control in a Single Server Queuing System (abstract) |
| 14:10 | Trimmed Minimum Error Entropy for Robust Online Regression (abstract) |
| 14:30 | May the force be with you (abstract) |
| 14:50 | Reachability of Chen-Fliess series: A Gradient Descent Approach (abstract) |
15:30-17:10 Session WeD1: Networks, Learning, and Algorithms III
Organizers: Bruce Hajek and R. Srikant
Chair:
Richard Sowers (University of Illinois at Urbana-Champaign, United States)
Location: Library
| 15:30 | Schedule Based Temporal Difference Algorithms (abstract) |
| 15:50 | A concentration bound for distributed stochastic approximation (abstract) |
| 16:10 | Global Convergence of Federated Learning for Mixed Regression (abstract) |
| 16:30 | Exploration, Exploitation, and Engagement in Multi-Armed Bandits with Abandonment (abstract) |
| 16:50 | On representation learning with Model-Agnostic Meta-Learning (abstract) |
15:30-17:10 Session WeD2: Robust Inference and Learning
Organizer: V. Veeravalli
Location: Solarium
| 15:30 | A distributionally robust approach to domain adaptation (abstract) |
| 15:50 | Policy Gradient Method For Robust Reinforcement Learning (abstract) |
| 16:10 | Post-hoc Uncertainty Learning using a Dirichlet Meta-Model (abstract) |
| 16:30 | Quickest Detection with Anticipatory and Quantum Decision Makers (abstract) |
| 16:50 | Policy Gradient based Entropic-VaR Optimization in Risk-Sensitive Reinforcement Learning (abstract) |
15:30-17:10 Session WeD3: Estimation, Prediction, and Control
Location: Butternut
| 15:30 | The Inverse Problem of Linear-Quadratic Differential Games: When is a Control Strategies Profile Nash? (abstract) |
| 15:50 | Adversarial Examples for Model-Based Control: A Sensitivity Analysis (abstract) |
| 16:10 | System Identification Entropy for Chen-Fliess Series and Their Interconnections (abstract) |
| 16:30 | Risk-Aware Safe Optimal Control of Uncertain Linear Systems (abstract) PRESENTER: Pouria Tooranjipour |
| 16:50 | Beyond UCB: statistical complexity and optimal algorithm for non-linear ridge bandits (abstract) |
15:30-17:10 Session WeD4: Queueing Systems
Location: Pine
| 15:30 | Queueing Delay Minimization in Overloaded Networks via Rate Control (abstract) |
| 15:50 | Age of Information Process under Strongly Mixing Communication -- Moment Bound, Mixing Rate and Strong Law (abstract) PRESENTER: Adrian Redder |
| 16:10 | ASUMAN: Age Sense Updating Multiple Access in Networks (abstract) |
| 16:30 | Universal Policy Tracking: Scheduling for Wireless Networks with Delayed State Observation (abstract) |
| 16:50 | Pricing, competition and market segmentation in ride hailing (abstract) |
Thursday, September 29th
View this program: with abstractssession overviewtalk overview
08:30-10:10 Session ThA1: Games and Distributed Optimization/Learning I
Organizers: Rasoul Etesami and Tamer Başar
Location: Library
| 08:30 | On the connection between reinforcement learning and gradient descent (abstract) |
| 08:50 | GRAND: A Gradient-Related Ascent and Descent Algorithmic Framework for Minimax Problems (abstract) |
| 09:10 | Random Adaptation Perspective to Distributed Computation (abstract) |
| 09:30 | On the Sample Complexity of Stabilizing LTI Systems on a Single Trajectory (abstract) |
08:30-10:10 Session ThA2: Electric Power Systems I
Organizers: A. Dominguez-Garcia and S. Bose
Chair:
Subhonmesh Bose (University of Illinois at Urbana-Champaign, United States)
Location: Solarium
| 08:30 | Convex Optimization of Bioprocesses (abstract) |
| 08:50 | A safe pricing algorithm for distributed demand management (abstract) |
| 09:10 | Storage Participation with State-of-Charge Dependent Offers and Bids: Market Clearing, Pricing, and Incentive Compatibility (abstract) |
| 09:30 | Inverse Power Flow Problem (abstract) |
| 09:50 | Synchronized Stop and Re-start Distributed Average Consensus for Control and Coordination Applications (abstract) |
08:30-10:10 Session ThA3: Intersection of Learning, Optimization, and Control III
Organizer: Bin Hu
Location: Butternut
| 08:30 | Unbounded Gradients in Federated Leaning with Buffered Asynchronous Aggregation (abstract) |
| 08:50 | Escaping saddle points in zeroth-order optimization: two function evaluations suffice (abstract) |
| 09:10 | Symmetric Natural Policy Gradient for Regularized Multi-Agent Learning with Parameter Convergence (abstract) |
| 09:30 | Black-box Control for Linear Dynamical Systems (abstract) |
08:30-10:10 Session ThA4: High-Dimensional Statistics and Learning II
Organizers: Bruce Hajek, Jiaming Xu, and Yihong Wu
Chair:
Feng Liang (University of Illinois at Urbana-Champaign, United States)
Location: Pine
| 08:30 | On the sample complexity of entropic optimal transport (abstract) |
| 08:50 | Detection-Recovery Gap for Planted Dense Cycles (abstract) |
| 09:10 | Recursive Causal Structure Learning (abstract) |
| 09:30 | The TAP free energy for high-dimensional linear regression with uniform spherical prior (abstract) |
| 09:50 | Local convexity of the TAP free energy and AMP convergence for Z2-synchronization (abstract) |
13:30-15:10 Session ThC1: High-Dimensional Statistics and Learning III
Organizers: Bruce Hajek, Jiaming Xu, and Yihong Wu
Chair:
Jiaming Xu (Duke University, United States)
Location: Library
| 13:30 | Phase transitions in the Gaussian database alignment and planted Gaussian matching problems (abstract) |
| 13:50 | Low-rank matrix estimation with groupwise heteroskedasticity (abstract) |
| 14:10 | Understanding the generalization power of overfitted NTK models: 3-layer vs. 2-layer (abstract) |
| 14:30 | Universality of High-Dimensional Estimation with Nearly Deterministic Sensing Matrices (abstract) |
13:30-15:10 Session ThC2: Electric Power Systems II
Organizers: A. Dominguez-Garcia and S. Bose
Location: Solarium
| 13:30 | Stable and Decentralized Learning for Voltage Regulation (abstract) |
| 13:50 | Model-Enhanced Localization of Forced Oscillation Using PMU Data (abstract) |
| 14:10 | Graph Convolutional Neural Networks for Control of Smart Inverters in Power Grids (abstract) |
| 14:30 | Bringing Data Science and Machine Learning into Electricity Markets (abstract) |
| 14:50 | Data-driven estimation of probabilistic constraints for network-safe distributed energy resource control (abstract) |
13:30-15:10 Session ThC3: Games and Networks
Location: Butternut
| 13:30 | Individual Altruism Cannot Overcome Congestion Effects in a Global Pandemic Game (abstract) |
| 13:50 | An Interactive Search Game with Two Agents (abstract) |
| 14:10 | Identifying Influential Agents Via Faux Adversarial Games (abstract) |
| 14:30 | Equilibrium characterizations of multi-resource Lotto games (abstract) |
| 14:50 | Equilibrium analysis of game on heterogeneous networks with coupled activities (abstract) |
13:30-15:10 Session ThC4: Control and Algorithms I
Organizer: Daniel Liberzon
Location: Pine
| 13:30 | Second order methods for min-max optimization with stability guarantees (abstract) |
| 13:50 | Rubik Tables, Stack Rearrangement, and Multi-Robot Routing (abstract) |
| 14:10 | Guaranteeing safety for systems with missing measurements (abstract) |
| 14:30 | Asymptotically Optimal Worst-Case State Estimation over Noisy Channels (abstract) |
| 14:50 | Proximal methods for self-healing and exact distributed convex optimization (abstract) |
15:30-17:10 Session ThD1: Electric Power Systems III
Organizers: A. Dominguez-Garcia and S. Bose
Location: Library
| 15:30 | Optimal Primary- and Secondary-control Design for Grids with Generators and Inverters (abstract) |
| 15:50 | Compensating Network Dynamics in Grid-Forming Control (abstract) |
| 16:10 | From Automatic Generation Control to Fast Frequency Control using Inverter-Based Resources (abstract) |
| 16:30 | Dynamic Performance of Unified Grid-Forming/Following Inverter Control (abstract) |
| 16:50 | Approaches to Analysis of Power System Dynamics with High Penetration of Inverter Interfaced Devices (abstract) |
15:30-17:30 Session ThD2: Games and Distributed Optimization/Learning II
Organizers: Rasoul Etasami and Tamer Başar
Location: Solarium
| 15:30 | Information provision to manage strategic agents over star networks - static and dynamic designs (abstract) |
| 15:50 | Gradient play in stochastic games: stationary points, convergence, and sample complexity (abstract) |
| 16:10 | Persuasion in Networks: Public Signals and Cores (abstract) |
| 16:30 | The Art of Concession in General Lotto Games (abstract) |
15:30-17:10 Session ThD3: Information Theory and Statistical Inference
Location: Butternut
| 15:30 | Minimum-Length Trace Reconstruction via Integer Programming (abstract) |
| 15:50 | Controlled Sensing with Corrupted Commands (abstract) |
| 16:10 | An Information-Theoretic Analysis of Bayesian Reinforcement Learning (abstract) |
| 16:30 | Finite-State Mutual Dimension (abstract) PRESENTER: Adam Case |
| 16:50 | On Learning a Hidden Directed Graph with Path Queries (abstract) |
15:30-16:30 Session ThD4: Theoretical Understandings of Machine Learning
Organizers: Yue M. Lu and Minh N. Do
Location: Pine
| 15:30 | Overparameterization improves robustness to covariate shift in high dimensions (abstract) |
| 15:50 | A Kernel Analysis of Feature Learning in Deep Neural Networks (abstract) |
16:40-17:20 Session ThD5: Control and Algorithms II
Location: Pine
| 16:40 | Opening the black box: deep-learned error-correcting codes (abstract) |
| 17:00 | On the Conley's decomposition for well-posed hybrid inclusions (abstract) |
Friday, September 30th
View this program: with abstractssession overviewtalk overview
08:30-10:10 Session FrA1: Networks, Learning, and Algorithms IV
Organizers: Bruce Hajek and R. Srikant
Chair:
Sasha Stolyar (University of Illinois at Urbana-Champaign, United States)
Location: Library
| 08:30 | Optimal Scheduling for Multi-Server Systems: Adversarial versus Stochastic Viewpoints (abstract) |
| 08:50 | Maximizing Utilization in Large Systems Serving Jobs with Time-Varying Resource Requirements (abstract) |
| 09:10 | Reward Poisoning Attacks on Offline Multi-Agent Reinforcement Learning (abstract) |
| 09:30 | Is Pessimism Provably Efficient for Offline RL? (abstract) |
| 09:50 | How to Charge Lightning: The Economics of Bitcoin Transaction Channels (abstract) |
08:30-09:50 Session FrA2: Multiuser Channels and Networks
Location: Solarium
| 08:30 | Achievable Rate Regions for the Gaussian Broadcast Channel with Fixed Blocklength and Per User Reliability (abstract) |
| 08:50 | Commitment over Multiple-Access Channels (abstract) |
| 09:10 | Throughput optimized random access channel assignment (abstract) |
| 09:30 | On the Limits of Distributed Agreement between Correlated Sources (abstract) |
| 09:50 | Multi-Message Private Information Retrieval: A Scalar Linear Solution (abstract) |
08:30-09:50 Session FrA3: Imaging and Data Science
Organizers: Zhizhen Zhao, Ivan Dokmanic, and Minh Do
Location: Butternut
| 08:30 | Multi-Pose Fusion for Sparse-View CT Reconstruction Using Consensus Equilibrium (abstract) PRESENTER: Diyu Yang |
| 08:50 | A Geometric Analysis of Neural Collapse with Unconstrained Features (abstract) |
| 09:10 | Robust Deep Image Prior with Partial Guidance (abstract) |
| 09:30 | A Regularized Conditional GAN for Posterior Sampling in Inverse Problems (abstract) |
10:10-11:30 Session FrB1: Networks, Learning, and Algorithms V
Organizers: Bruce Hajek and R. Srikant
Chair:
Vijay Subramanian (University of Michigan, United States)
Location: Library
| 10:10 | Bias in Stochastic Approximation Cannot Be Eliminated With Averaging (abstract) |
| 10:30 | Practical Control Design for the Deep Learning Age: Distillation of Deep RL-Based Controllers (abstract) |
| 10:50 | Learning Preference Distributions From Distance Measurements (abstract) |
| 11:10 | Ignoring Causality to Improve Ranking (abstract) |
10:10-11:30 Session FrB2: Topics in Optimization
Location: Solarium
| 10:10 | Optimal Discrimination Between Two Pure States and Dolinar-Type Coherent-State Detection (abstract) |
| 10:30 | Alternating Direction Method of Multipliers for Decomposable Saddle-Point Problems (abstract) |
| 10:50 | Q-linear Convergence of Distributed Optimization with Barzilai-Borwein Step Sizes (abstract) |
| 11:10 | Distributed Online Non-convex Optimization with Composite Regret (abstract) |
| 11:30 | The Effects of Varying Charging Rates on Optimal Charging Station Choices for Electric Vehicles (abstract) |
10:10-11:30 Session FrB3: Coding Techniques
Location: Butternut
| 10:10 | Covering Properties of Sum-Rank Metric Codes (abstract) |
| 10:30 | Hardness of Linear Index Coding on Perturbed Instances (abstract) |
| 10:50 | Degree Tables for Private Information Retrieval (abstract) |
| 11:10 | Secure Linear MDS Coded Matrix Inversion (abstract) |
| 11:30 | Grassmannian packings: Trust region stochastic tuning for matrix incoherence (abstract) PRESENTER: Josiah Park |
| 11:50 | An Error Exponent for the AWGN Channel with Decision Feedback and Lattice Coding (abstract) |