Publications


Forthcoming

DeepSplit: Scalable Verification of Deep Neural Networks via Operator Splitting (arXiv)
Shaoru Chen, Eric Wong, J Zico Kolter, Mahyar Fazlyab

On Centralized and Distributed Mirror Descent: Exponential Convergence Analysis Using Quadratic Constraints (arXiv)
Youbang Sun, Mahyar Fazlyab, Shahin Shahrampour


2022

Safety verification and robustness analysis of neural networks via quadratic constraints and semidefinite programming (arXiv)
Mahyar Fazlyab, Manfred Morari, George J Pappas
IEEE Transactions on Automatic Control, 2020


2021

Learning region of attraction of nonlinear systems (arXiv)
Shaoru Chen, Mahyar Fazlyab, Manfred Morari, George J Pappas, Victor M Preciado
American Control Conference, 2021

Learning Lyapunov functions for hybrid systems
Shaoru Chen, Mahyar Fazlyab, Manfred Morari, George J Pappas, Victor M Preciado
International Conference on Hybrid Systems: Computation and Control, 2021

Stability analysis of complementarity systems with neural network controllers (arXiv)
Alp Aydinoglu, Mahyar Fazlyab, Manfred Morari, Michael Posa
International Conference on Hybrid Systems: Computation and Control, 2021

Performance Bounds for Neural Network Estimators: Applications in Fault Detection (arXiv)
Navid Hashemi, Mahyar Fazlyab, Justin Ruths
American Control Conference, 2021

Enforcing robust control guarantees within neural network policies (arXiv)
Priya L Donti, Melrose Roderick, Mahyar Fazlyab, J Zico Kolter
International Conference on Learning Representations, 2021

Certifying Incremental Quadratic Constraints for Neural Networks
via Convex Optimization
(arXiv)
Navid Hashemi, Justin Ruths, Mahyar Fazlyab
Learning for Dynamics & Control Conference (L4DC), 2021


2020

Reach-SDP: Reachability analysis of closed-loop systems with neural network controllers via
semidefinite programming
(arXiv)
Haimin Hu, Mahyar Fazlyab, Manfred Morari, George J Pappas
IEEE Conference on Decision and Control (CDC), 2020

Robust deep learning as optimal control: Insights and convergence guarantees (arXiv)
Jacob H Seidman, Mahyar Fazlyab, Victor M Preciado, George J Pappas
Learning for Dynamics & Control Conference (L4DC), 2020

Network statistics estimation and prediction
Mahyar Fazlyab, Mehdi Nikkhah, Mark Allen Webb
US Patent


2019

Probabilistic verification and reachability analysis of neural networks via semidefinite programming (arXiv)
Mahyar Fazlyab, Manfred Morari, George J Pappas
IEEE Conference on Decision and Control (CDC), 2019

Robust convergence analysis of Three-Operator Splitting (arXiv)
Han Wang, Mahyar Fazlyab, Shaoru Chen, Victor M Preciado
Annual Allerton Conference on Communication, Control, and Computing (Allerton), 2019

A Prediction-Correction Primal-Dual Algorithm for Distributed Optimization
Santiago Paternain, Mahyar Fazlyab, Victor M Preciado, Alejandro Ribeiro
2019 American Control Conference (ACC), 2019

Efficient and accurate estimation of Lipschitz constants for deep neural networks (arXiv)
Mahyar Fazlyab, Alexander Robey, Hamed Hassani, Manfred Morari, George Pappas
Advances in Neural Information Processing Systems (NeurIPS), 2019

A control-theoretic approach to analysis and parameter selection of Douglas–Rachford splitting (arXiv)
Jacob H Seidman, Mahyar Fazlyab, Victor M Preciado, George J Pappas
IEEE Control Systems Letters, 2019


2018

A Chebyshev-accelerated primal-dual method for distributed optimization (arXiv)
Jacob H Seidman, Mahyar Fazlyab, George J Pappas, Victor M Preciado
IEEE Conference on Decision and Control (CDC), 2018

Design of first-order optimization algorithms via sum-of-squares programming (arXiv)
Mahyar Fazlyab, Manfred Morari, Victor M. Preciado
IEEE Conference on Decision and Control (CDC), 2018

Distributed smooth and strongly convex optimization with inexact dual methods
Mahyar Fazlyab, Santiago Paternain, Alejandro Ribeiro, Victor M Preciado
Annual American Control Conference (ACC), 2018

Analysis of optimization algorithms via integral quadratic constraints: Nonstrongly convex problems (arXiv)
Mahyar Fazlyab, Alejandro Ribeiro, Manfred Morari, Victor M Preciado
SIAM Journal on Optimization, 2018

Prediction-correction interior-point method for time-varying convex optimization (arXiv)
Mahyar Fazlyab, Santiago Paternain, Victor M Preciado, Alejandro Ribeiro
IEEE Transactions on Automatic Control, 2018


2017

A dynamical systems perspective to convergence rate analysis of proximal algorithms
Mahyar Fazlyab, Alejandro Ribeiro, Manfred Morari, Victor M Preciado
Annual Allerton Conference on Communication, Control, and Computing (Allerton), 2017

Optimal network design for synchronization of coupled oscillators (arXiv)
Mahyar Fazlyab, Florian Dörfler, Victor M Preciado
Automatica, 2017

A variational approach to dual methods for constrained convex optimization
Mahyar Fazlyab, Alec Koppel, Victor M Preciado, Alejandro Ribeiro
American Control Conference (ACC), 2017


2016

Self-triggered time-varying convex optimization (arXiv)
Mahyar Fazlyab, Cameron Nowzari, George J Pappas, Alejandro Ribeiro, Victor M Preciado
55th Conference on Decision and Control (CDC), 2016

Interior point method for dynamic constrained optimization in continuous time (arXiv)
Mahyar Fazlyab, Santiago Paternain, Victor M Preciado, Alejandro Ribeiro
American Control Conference (ACC), 2016

2015 and Prior

Rotation rate estimation in parametrically excited micro gyroscopes
Mahyar Fazlyab, Hassan Salarieh, Aria Alasty
Mechatronics, 2015

Robust topology identification and control of LTI networks
Mahyar Fazlyab, Victor M Preciado
IEEE Global Conference on Signal and Information Processing (GlobalSIP), 2014

Parameter estimation and interval type-2 fuzzy sliding mode control of a z-axis MEMS gyroscope
Mahyar Fazlyab, Maysam Zamani Pedram, Hassan Salarieh, Aria Alasty
ISA transactions, 2013