Responsive image

Research Interests

I am interested in understanding the role of information in monitoring, communication, and decision making systems. My research lies broadly in:
  • Information theory and data science
  • Machine learning and high-dimensional statistics
  • Cybersecurity
    • Detection and mitigation of data integrity attacks
    • Privacy and confidentiality
  • Information processes on cyberphysical systems
    • Robust estimation with imperfect information
    • Optimal sensor placement problems

News

  • October 28, 2024: Our paper "Information-Theoretic Sensor Placement for Large Sewer Networks" has been accepted for publication in Water Research. This is George's first journal publication! Joint work with S. Tait, G. Panoutsos, and V. Speight. [Preprint]
  • September 5, 2024: PhD student George Crowley has proved the submodularity of mutual information for Gaussian multivariate sources observed under additive white Gaussian noise. This is a very useful result for sensor placement and observer design problems. The main result is described in the technical report [here]
  • July 24, 2024: We will be presenting a tutorial on "Characterizing the Generalization Error of Machine Learning Algorithms Via Information Measures" at ITW 24 to be held in Shenzhen, China. This is joint work with G. Aminian, Y. Bu, and S.M. Perlaza. We wil publish a more detailed outline but a brief tutorial descripition can be found [here].
  • April 09, 2024: PhD student Francisco Daunas' paper "Equivalence of the Empirical Risk Minimization to Regularization on the Family of f-Divergences" has been accepted for presentation at ISIT 24. This is joint work with S.M. Perlaza and H.V. Poor. [Preprint]
  • March 25, 2024: Our paper "The Worst-Case Data-Generating Probability Measure in Statistical Learning" has been accepted to IEEE Journal on Selected Areas in Information Theory. This is joint work with X. Zou, S.M. Perlaza, E. Altman, and H.V. Poor. [Preprint]
  • February 20, 2024: Our paper "An Information Theoretic Metric for Measurement Vulnerability to Data Integrity Attacks on Smart Grids" has been accepted for publication in IET Smart Grid. This is joint work with Xiuzhen Ye, S.M. Perlaza, and R.F. Harrison. [Preprint]
  • January 30, 2024: Our paper "Empirical Risk Minimization with Relative Entropy Regularization" has been accepted for publication in IEEE Transactions on Information Theory. This is joint work with S.M. Perlaza, G. Bisson, A. Jean-Marie, and S. Rini. [Preprint]
  • January 11, 2024: Our paper "Stability of Nonlinear Model Predictive Control under Consecutive Packet Losses" has been accepted to UKACC 2024 to be held in University of Southampton. This is joint work with Paulo Loma and Paul Trodden.
  • December 9, 2023: Our paper "Generalization Analysis of Machine Learning Algorithms via the Worst-Case Data-Generating Probability Measure" has been accepted to AAAI-2024 in Vancouver. This is joint work with X. Zou (first conference paper for Xinying!), S.M. Perlaza, and E. Altman. [Preprint]
  • November 28, 2023: I will be giving a talk in the University of Cambridge, Information Theory Seminar, about our recent work on machine learning regularization with information measures. [Details]
  • April 14, 2023: We've had two papers accepted for presentation at ISIT 23, one of them is the first paper for our PhD student Francisco Daunas! Preprints of the papers here:
    • F. Daunas, I. Esnaola, S.M. Perlaza, and H.V. Poor, "Analysis of the Relative Entropy Asymmetry in the Regularization of Empirical Risk Minimization". [Preprint]
    • S.M. Perlaza, I. Esnaola, G. Bisson, and H.V. Poor, "On the Validation of Gibbs Algorithms: Training Datasets, Test Datasets and their Aggregation". [Preprint]
  • January 30, 2023: Our paper "Stealth Data Injection Attacks with Sparsity Constraints" has been accepted for publication in IEEE Transactions on Smart Grid. This is joint work with X. Ye, S.M. Perlaza, and R.F. Harrison. [Preprint]
  • January 13, 2023: Our paper "Asymptotic Learning Requirements for Stealth Attacks on Linearized State Estimation" has been accepted for publication in IEEE Transactions on Smart Grid. This is joint work with K. Sun, A.M. Tulino, and H.V. Poor. [Preprint]