Causal Machine Learning
When observational data contain unobserved confounding variables, it is well known that the causal effects of these confounders cannot be estimated without bias regardless of the sample size. As a result, any associations between variables that are learned from the data may be spurious or distorted leading to false predictions. In this research, we rely on rigorous principles of causal inference to develop new machine learning methods that can better explain and predict the causal effects of policies and interventions as well as control for biases in the data. Applications include robotics and healthcare. Research supported by NSF and AFOSR.
Y. Shen, J. Dunn, and M. M. Zavlanos, "Risk-Averse Multi-Armed Bandits with Unobserved Confounders: A Case Study in Emotion Regulation in Mobile Health," Proc. 61st IEEE Conference on Decision and Control (CDC), Cancun, Mexico, Dec. 2022.
C. Liu, Y. Zhang, Y. Shen, and M. M. Zavlanos, "Learning without Knowing: Unobserved Context in Continuous Transfer Reinforcement Learning," Proc. 3rd Conference on Learning for Dynamics and Control (L4DC), ser. Proc. of Machine Learning Research, vol. 144, pp. 791-802, Jun. 2021.
Optimization for Machine Learning and AI
Optimization algorithms are at the core of machine learning and AI. In fact, many machine learning and AI models can be viewed as solutions to appropriate optimization problems. In this research, we develop new optimization algorithms for machine learning and AI that can handle distributed agents and data, distribution shifts in the data, and non-stationary environments, and analyze their convergence and complexity properties. Applications include robotics and healthcare. Research supported by NSF and AFOSR.
Y. Shen, P. Xu, and M. M. Zavlanos, "Wasserstein Distributionally Robust Policy Evaluation and Learning for Contextual Bandits," under review.
Y. Zhang and M. M. Zavlanos, "Cooperative Multi-Agent Reinforcement Learning with Partial Observations," IEEE Transactions on Automatic Control, Jun. 2023, DOI: 10.1109/TAC.2023. 3288025
Black-Box Optimization and Learning
Zeroth-order (or derivative-free) optimization methods enable the optimization of black-box models that are available only in the form of input-output data and are common in simulation-based optimization, training of Deep Neural Networks, and reinforcement learning. In the absence of input-output models, exact first or second order information (gradient or hessian) is unavailable and cannot be used for optimization. Therefore, zeroth-order methods rely on input-output data to obtain approximations of the gradients that can be used as descent directions. In this research, we develop new zeroth-order algorithms for distributed and non-stationary optimization and learning problems with reduced variance and improved complexity. Applications include robotics and healthcare. Research supported by NSF and AFOSR.
Z. Wang, X. Yi, Y. Shen, M. M. Zavlanos, and K. H. Johansson, "Asymmetric Feedback Learning in Online Convex Games," under review.
Y. Zhang, Y. Zhou, K. Ji, and M. M. Zavlanos, "A New One-Point Residual-Feedback Oracle for Black-Box Learning and Control," Automatica, vol. 136, pp. 110006, Feb. 2022.
Safe Learning for Control
Recent progress in the field of machine learning has given rise to a new family of neural network controllers for robotic systems that can significantly simplify the overall design process. As such control schemes become more common in real-world applications, the ability to train neural networks with safety considerations becomes a necessity. In this research, we develop new safe learning methods for robot navigation problems and study the tradeoff between data density, computational complexity, and safety guarantees. Research supported by AFOSR.
P. Vlantis, L. J. Bridgeman, and M. M. Zavlanos, "Failing with Grace: Learning Neural Net- work Controllers that are Boundedly Unsafe," Proc. 5th Conference on Learning for Dynamics and Control (L4DC), ser. Proc. of Machine Learning Research (PMLR), N. Matni, M. Morari, and G. J. Pappas, Eds., vol. 211, pp. 954-965, Jun. 2023.
S. Sun, Y. Zhang, X. Luo, P. Vlantis, M. Pajic, and M. M. Zavlanos, "Formal Verification of Stochastic Systems with ReLU Neural Network Controllers," Proc. 2022 International Conference on Robotics and Automation (ICRA), Philadelphia, PA, May 2022, pp. 6800-6806.