Risk-Aware Optimization for Statistical Estimation and Control

Time: Wednesday, February 24, 2021 - 11:30am - 12:30pm
Type: Seminar Series
Presenter: Dionysios Kalogerias; Michigan State University
Room/Office:
Location:

Department of Electrical Engineering Seminar

Dionysios Kalogerias
"Risk-Aware Optimization for Statistical Estimation and Control"

Wednesday, February 24, 2021; 11:30 AM
Zoom Presentation (contact dept. for login details)

Abstract: Modern, critical applications require that stochastic decisions for estimation and control are made not only on the basis of minimizing average losses, but also safeguarding against less frequent, though possibly catastrophic events. Examples appear naturally in many areas, such as energy, finance, robotics, radar/lidar, networking and communications, autonomy, safety, and the Internet-of-Things. In such applications, the ultimate goal is to obtain risk-aware decision policies that optimally compensate against extreme events, even at the cost of slightly sacrificing performance under nominal conditions.

In the first part of this talk, we present and discuss a new risk-aware formulation of the classical and ubiquitous nonlinear MMSE estimation problem, trading between mean performance and risk by explicitly constraining the expected predictive variance of the squared error. Quite remarkably, we show that the optimal risk-aware solution can be evaluated stably and in closed-form regardless of the underlying generative model, as an appropriately biased, interpolated novel nonlinear MMSE estimator with a rational structure. We further illustrate the effectiveness of our approach via numerical examples, showcasing the advantages of risk-aware MMSE estimation against risk-neutral MMSE estimation, especially in models involving skewed and/or heavy-tailed distributions.

We then turn our attention to the stochastic LQR control paradigm. Driven by the ineffectiveness of risk-neutral LQR controllers at the presence of risky events, we present a new risk-constrained LQR formulation, which restricts the total expected predictive variance of the state penalty by a user-prescribed level. Again, the optimal controller can be evaluated in closed form. In fact, it is affine relative to the state, internally stable regardless of parameter tuning, and strictly optimal under minimal assumptions on the process noise (i.e., finite fourth-order moments), effectively resolving the shortcomings of Linear-Exponential-Gaussian (LEG) control put forward by David Jacobson and Peter Whittle in the 1970-80's, also being the current state of the art. The advertised advantages of the new risk-aware LQR framework are further illustrated via indicative numerical examples.

Bio: Dionysios Kalogerias received the MEng degree from the Department of Computer Engineering and Informatics, University of Patras, Greece (2010), the MSc degree in signal processing and communications through the interdepartmental graduate program SPCOMS, University of Patras, Greece (2012), and the PhD degree with double distinction from the Department of Electrical and Computer Engineering (ECE), Rutgers, The State University of New Jersey, Piscataway, NJ (2017). Dionysios is currently an assistant professor with the Department of ECE, Michigan State University (MSU). Prior to joining MSU, he was a postdoctoral researcher with the Department of Electrical and Systems Engineering, University of Pennsylvania, Philadelphia, PA, where he was affiliated with GRASP, Alelab, and PRECISE (2019 – 2020). Prior to that, he was a postdoctoral research associate with the Department of Operations Research and Financial Engineering, Princeton University, Princeton, NJ (2017 – 2019), where he was affiliated with CASTLE Labs. His research interests lie in the areas of machine learning, reinforcement learning, optimization, signal processing, sequential decision making, and risk, and their applications in autonomous networked systems, wireless communications, security and privacy, and system trustworthiness.

Dionysios has been the recipient of several awards from Rutgers, including a Leeds Fellowship, the ECE PhD Student Research Excellence Award, and the School of Engineering Professional Development Fund Award. For his doctoral thesis, he received both the 2017 Rutgers ECE Graduate Program Academic Achievement Award, and the 2017 Rutgers School of Engineering Outstanding Graduate Student Award. He has also received the ICASSP 2016 Best Student Paper of the Special Sessions Award (also nominated for a Best Student Paper Award) and, more lately, he was the sole recipient of the ICASSP 2020 Best Paper Award.

Hosted by: Prof. Amin Karbasi