**
Communication, Control**
**
and Signal Processing Seminar****
**

**Abstracts
of talks**

*Fall 2019*

** 09/19 Yasser Shoukry (UMD)****,
Secure State Estimation: A Satisfiability Modulo Convex Programming Approach
**

ABSTRACT: The rapidly increasing dependence on Cyber-Physical Systems (CPS) in building critical infrastructures—in the context of smart cities, power grids, medical devices, and self-driving cars—has opened the gates to increasingly sophisticated and harmful attacks with financial, societal, criminal or political effects. While a traditional cyber-attack may leak credit-card or other personal sensitive information, a CPS-attack can lead to a loss of control in nuclear reactors, gas turbines, the power grid, transportation networks, and other critical infrastructure, placing the Nation’s security, economy, and public safety at risk. In this talk, I will focus on one the problem of estimating the state of a dynamical system when an adversary arbitrarily corrupts a subset of its sensors. Although of critical importance, this problem is NP-hard and combinatorial since the subset of attacked sensors is unknown. I will show how to tame the combinatorial nature of the problem using a novel technique called "Satisfiability Modulo Convex Programming" or SMC for short.

** 09/26 Alec Koppel (U.S. Army Research Laboratory)****,
Policy Search for Reinforcement Learning in Continuous Spaces: Improved Limits and Reduced Variance
**

ABSTRACT: Reinforcement Learning (RL) is a form of stochastic adaptive control in which one seeks to estimate parameters of a controller without having access to a dynamics model. RL has gained popularity in recent years beginning with the smashing success of AlphaGo besting the world champion in Go during summer 2016. However, the recent empirical successes of RL have been called into question due to their irreproducibility and high variance across different training runs. Motivated by this gap, we'll spotlight recent efforts to solidify theoretical understanding of the rate analysis and limiting properties of policy gradient methods in continuous Markov Decision Problems from a non-convex optimization perspective. Moreover, we design modified step-size rules that yield convergence to approximate local extrema, motivating reward-reshaping via nonconvex optimization. We'll then discuss a modification of the Policy Gradient Theorem that yields provably lower variance policy search directions, and algorithms based upon which yield algorithms with reduced variance. These results provide a conceptual framework for the future design of stable RL tools with lower variance.

** 11/07 Xiaodi Wu (UMD)****,
Quantum query complexity of entropy estimation
**

ABSTRACT: Estimation of Shannon and R{'e}nyi entropies of unknown discrete distributions is a fundamental problem in statistical property testing and an active research topic in both theoretical computer science and information theory. Tight bounds of the number of samples to estimate these entropies have been established in the classical setting, while little is known about their quantum counterparts. In this talk, I will show quantum algorithms for estimating $\alpha$-R{'e}nyi entropies (Shannon entropy being 1-Renyi entropy). In particular, I will demonstrate a quadratic quantum speedup for Shannon entropy estimation and a generic quantum speedup for $\alpha$-R{'e}nyi entropy estimation for all $\alpha$>0, including a tight bound for the collision-entropy (2-R{'e}nyi entropy) and also an analysis for the min-entropy case (i.e., $\alpha$ = +infinity). This talk is based on joint work with Tongyang Li.

** 11/21 Behtash Babadi (UMD)****,
Dynamic Network-level Analysis of Neural Data Underlying Behavior: Beyond the Linear, Static, and Gaussian Domains
**

ABSTRACT: In this talk, I present computational methodologies for extracting dynamic neural functional networks that underlie behavior. These methods aim at capturing the sparsity, dynamicity and stochasticity of these networks, by integrating techniques from high-dimensional statistics, point processes, state-space modeling, and adaptive filtering. I demonstrate their utility using several case studies involving auditory processing, including 1) functional auditory-prefrontal interactions during attentive behavior in the ferret brain, 2) network-level signatures of decision-making in the mouse primary auditory cortex, and 3) cortical dynamics of speech processing in the human brain.

** 12/05 Benjamin Kedem (UMD)****,
Estimation of Small Tail Probabilities by Repeated out of Sample Fusion
**

ABSTRACT: [pdf]