Abstract: Future applications of national importance, such as healthcare, critical infrastructure, transportation systems, and smart cities, are expected to increasingly rely on machine-learning methods, including structured learning, supervised learning, and reinforcement learning. In many of these applications, the probabilistic distribution governing the data may undergo variations with time and location, and data could be corrupted by faulty or malicious agents/sensors. Such model deviation and data corruption could result in significant performance degradation. The goal in this project is to explore new ways to design learning and inference methods that are robust to distributional uncertainty and data corruption. This project is bridging and further advancing research in areas of statistical learning, optimization, control theory, network science, reinforcement learning, statistical signal processing and information theory. The methods developed are likely to have significant impact on a wide range of applications in areas of societal importance such as healthcare, transportation systems, smart grids, and smart cities. The investigators are co-organizing special sessions at conferences, workshops and symposia on robust learning and inference to disseminate the research outcomes of this project, formalize far-reaching research directions, identify new challenges in this emerging area, stimulate the development of original research ideas, and foster interdisciplinary collaborations. The investigators are committed to broadening participation of under-represented minorities and women both among the graduate and undergraduate students in computing and engineering. The investigators are enriching their current courses and further developing new courses on topics related to this project.
This project is expected to make new contributions to the theory and practice of robust learning and inference. Several emerging directions are being investigated, including robust sketch-based learning, robust mean estimation, synthesis of confusing inputs to machine-learning models, robustness to distributional uncertainty at inference time, and robust model-free reinforcement learning.
Selected publications:
Model-Free Robust Average-Reward Reinforcement Learning
Y. Wang, A. Velasquez, G. Atia, A. Prater-Bennette, S. Zou
International Conference on Machine Learning (ICML), 2023Robust Low-tubal-rank Tensor Completion based on Tensor Factorization and Maximum Correntopy Criterion
Y. He and G. Atia
IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 2023Kernel Robust Hypothesis Testing
Z. Sun, S. Zou
IEEE Transactions on Information Theory, 2023Robust Average-Reward Markov Decision Processes
Y. Wang, A. Velasquez, G. Atia, A. Prater-Bennette, S. Zou
Thirty-Seventh AAAI Conference on Artificial Intelligence (AAAI), 2023, Distinguished Paper AwardPatch Tracking-based Streaming Tensor Ring Completion for Visual Data Recovery
Y. He and G. Atia
IEEE Transactions on Circuits and Systems for Video Technology (TCSVT), Dec 2022.A Differentiable Approach to the Maximum Independent Set Problem using Dataless Neural Networks
I. Alkhouri, A. Velasquez and G. Atia
Neural Networks, Elsevier, Nov 2022Sketch-based community detection in evolving networks
A. Beckus and G. Atia
Phys. Rev. E, 106, 044306, Oct 2022Coarse to Fine Two-Stage Approach to Robust Tensor Completion of Visual Data
Y. He and G. Atia
IEEE Transactions on Cybernetics, Sep 2022Data-Driven Robust Multi-Agent Reinforcement Learning
Y. Wang, Y. Wang, Y. Zhou, A. Velasquez, S. Zou
IEEE International Workshop on Machine Learning for Signal Processing, 2022On the Coarse Robustness of Classifiers
I. Alkhouri, A. Velasquez, S. Bak, and G. Atia
Asilomar Conference on Signals, Systems, and Computers, Oct. 2022BOSS: Bidirectional One-Shot Synthesis of Adversarial Examples
I. Alkhouri, G. Atia and A. Velasquez
IEEE International Workshop on Machine Learning for Signal Processing (MLSP), 2022The Minimum Value State Problem in Actor-Critic Networks
A. Velasquez, B. Bissey, I. Alkhouri, L. Barak, G. Atia
IEEE International Workshop on Machine Learning for Signal Processing (MLSP), 2022An Approach to the Maximum Independent Set Problem Using Graph-based Neural Network Structures
I. Alkhouri, G. Atia and A. Velasquez
IEEE International Workshop on Machine Learning for Signal Processing (MLSP), 2022Policy Gradient Method For Robust Reinforcement Learning
Y. Wang, S. Zou
International Conference on Machine Learning (ICML), 2022.Sample and Communication-Efficient Decentralized Actor-Critic Algorithms with Finite-Time Analysis
Z. Chen, Y. Zhou, R. Chen, S. Zou
International Conference on Machine Learning (ICML), 2022.Robust Mean Estimation in High Dimensions: An Outlier-Fraction Agnostic and Efficient Algorithm
A. Deshmukh, J. Liu and V.V. Veeravalli
IEEE International Symposium on Information Theory (ISIT), 2022.Robust Hypothesis Testing With Kernel Uncertainty Sets
Z. Sun, S. Zou
IEEE International Symposium on Information Theory (ISIT), 2022.Active Grammatical Inference for Non-Markovian Planning
N. Topper, G. Atia, A. Trivedi and A. Velasquez
32nd International Conference on Automated Planning and Scheduling (ICAPS), June 2022.Inferring Probabilistic Reward Machines from Non-Markovian Reward Signals for Reinforcement Learning
T. Dohmen, N. Topper, G. Atia, A. Beckus, A. Trivedi and A. Velasquez
32nd International Conference on Automated Planning and Scheduling (ICAPS), June 2022.Multi-Agent Tree Search with Dynamic Reward Shaping
A. Velasquez, B. Bissey, L. Barak, D. Melcer, A. Beckus, I. Alkhouri and G. Atia
32nd International Conference on Automated Planning and Scheduling (ICAPS), June 2022.Synthesis of Adversarial Samples in Two-Stage Classifiers
I. Alkhouri, A. Velasquez and G. Atia
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Singapore, May 2022.Controller Synthesis for Omega-Regular and Steady-State Specifications
A. Velasquez, I. Alkhouri, A. Beckus, A. Trivedi and G. Atia
International Conference on Autonomous Agents and Multiagent Systems (AAMAS), May 2022.Multi-mode tensor space clustering based on low-tensor-rank representation
Y. He, G. Atia
Thirty-Sixth Conference on Artificial Intelligence (AAAI), Virtual, Feb 2022.Online Robust Reinforcement Learning with Model Uncertainty
Y. Wang, S. Zou
Conference on Neural Information Processing Systems (NeurIPS), Virtual, Dec 2021.Non-Asymptotic Analysis for Two Time-scale TDC with General Smooth Function Approximation
Y. Wang, Y. Zhou, S. Zou
Conference on Neural Information Processing Systems (NeurIPS), Virtual, Dec 2021.Sketches by MoSSaRT: Representative Selection from Manifolds with Gross Sparse Corruptions
M. Sedghi, M. Georgiopoulos, G. Atia
Pattern Recognition, Accepted Nov 2021.Steady-State Planning in Expected Reward Multichain MDPs
G. Atia, A. Beckus, I. Alkhouri, A. Velasquez
Journal of Artificial Intelligence Research (JAIR), 72 (2021) 1029-1082High-Dimensional Robust Mean Estimation via Outlier-Sparsity Minimization
A. Deshmukh, J. Liu, and V. V. Veeravalli.
IEEE Asilomar Conference on Signals, Systems and Computers, PacificGrove, CA, (virtual), November 2021.
Scalable Community Detection in the Degree-Corrected Stochastic Block Model
Y. He, A. Beckus, G. Atia
IEEE International Workshop on Machine Learning for Signal Processing (MLSP), Virtual, Oct 2021.Adversarial Attacks on Multi-level Fault Detection and Diagnosis Systems
A. Awad, I. Alkhouri, G. Atia
IEEE International Workshop on Machine Learning for Signal Processing (MLSP), Virtual, Oct 2021.Adversarial Perturbation Attacks on Nested Dichotomies Classification Systems
I. Alkhouri, A. Velasquez, G. Atia
IEEE International Workshop on Machine Learning for Signal Processing (MLSP), Virtual, Oct 2021.