Nikolai Matni 

Nikolai Matni
Postdoctoral Scholar
Electrical Engineering and Computer Science
UC Berkeley

Research Interests

Machine and Reinforcement Learning, Robust and Distributed Optimal Control, Convex Optimization, Cyber-Physical Systems, Software Defined Networking

Research Overview

My work aims to develop foundational theory and computational tools, rooted in machine learning, optimization, and control for the design of large-scale data-driven cyber-physical systems (DDCPS) such as the smart-grid, software-defined networks and smart transportation systems. For such a theory to be successful, it must provide the non-asymptotic guarantees of contemporary high-dimensional statistics, the stability, safety and performance guarantees of robust control theory, and the ability to learn and adapt of data-driven systems, all while being applicable to safety-critical systems that are highly dynamic and interconnected, difficult to model, ever changing and of huge scale. The goal of my research is to develop such a theoretical framework and the corresponding computational tools needed to make these insights actionable. Please see my publications page for current research projects, and the talk below for an accessible introduction (aimed at a general audience) to some of the ideas behind my work.

Recent Travel and Updates

  • Apr 5: EE Seminar at USC

  • Mar 14: ME Seminar at UCSB

  • Feb 26: ECE Seminar at the University of Minnesota

  • Feb 15: ECE Seminar at UIUC

  • Jan 29: EE Seminar at UCLA

  • Jan 22: ECE Seminar at the University of Michigan

  • I will be serving on the program committee for the SIGCOMM’18 workshop “Self-DN - Self-Driving Networks”

  • Dec 12-15: IEEE CDC 2017 in Melbourne, Australia

  • Dec 6-8: The ACEMS Workshop on Challenges of Data and Control of Networks (ACDCN) at the University of Adelaide