Mario Diaz

Mario Diaz

Research Associate
Instituto de Investigaciones en Matemáticas Aplicadas y en Sistemas (IIMAS)
Universidad Nacional Autónoma de México (UNAM)
Google Scholar Profile / WOS Profile / Curriculum Vitae

Interests

My research interests lie on the intersection of information theory, probability, and statistics.
My work mainly focuses on data privacy, theoretical machine learning, and random matrices.

Publications and Preprints

Journal Papers

  1. M. Diaz and J. Mingo. On the analytic structure of second-order non-commutative probability spaces and functions of bounded Fréchet variation.
    Random Matrices: Theory and Applications, 2023. DOI:10.1142/S2010326322500447
     
  2. M. Diaz, A. Jaramillo and J.C. Pardo. Fluctuations for matrix-valued Gaussian processes.
    Annales de l’Institut Henri Poincaré (B) Probabilités et Statistiques, 2022. DOI:10.1214/21-AIHP1238
     
  3. T. Sypherd, M. Diaz, J. Cava, G. Dasarathy, P. Kairouz and L. Sankar. A tunable loss function for robust classification: calibration, landscape, and generalization.
    IEEE Transactions on Information Theory, 2022. DOI:10.1109/TIT.2022.3169440
     
  4. H. Wang, H. Hsu, M. Diaz and F. Calmon. To split or not to split: the impact of disparate treatment in classification.
    IEEE Transactions on Information Theory, 2021. DOI:10.1109/TIT.2021.3075415
     
  5. M. Diaz, H. Wang, F. Calmon and L. Sankar. On the robustness of information-theoretic privacy measures and mechanisms.
    IEEE Transactions on Information Theory, 2020. DOI:10.1109/TIT.2019.2939472
     
  6. M. Diaz, J. Mingo and S. Belinschi. On the global fluctuations of block Gaussian matrices.
    Probability Theory and Related Fields, 2020. DOI:10.1007/s00440-019-00925-1
     
  7. S. Asoodeh, M. Diaz, F. Alajaji and T. Linder. Estimation efficiency under privacy constraints.
    IEEE Transactions on Information Theory, 2019. DOI:10.1109/TIT.2018.2865558
     
  8. M. Diaz and V. Pérez-Abreu. On the capacity of block multiantenna channels.
    IEEE Transactions on Information Theory, 2017. DOI:10.1109/TIT.2017.2712711
     
  9. S. Asoodeh, M. Diaz, F. Alajaji and T. Linder. Information extraction under privacy constraints.
    Information, 2016. DOI:10.3390/info7010015

Preprints

  1. J. Perusquía, M. Diaz and R. Mena. On a divergence-based prior analysis of stick-breaking processes. arXiv:2308.11868
     
  2. S. Asoodeh and M. Diaz. Privacy loss of noisy stochastic gradient descent might converge even for non-convex losses. arXiv:2305.09903
     
  3. M. Diaz, P. Kairouz and L. Sankar. Lower bounds for the MMSE via neural network estimation and their applications to privacy. arXiv:2108.12851
     
  4. S. Asoodeh, M. Diaz and F. Calmon. Privacy analysis of online learning algorithms via contraction coefficients. arXiv:2012.11035

Conference Papers

  1. S. Asoodeh and M. Diaz. On the privacy guarantees of differentially private stochastic gradient descent.
    Accepted for presentation at the IEEE International Symposium on Information Theory (ISIT), 2024.
     
  2. B. Zamanlooy, S. Asoodeh, M. Diaz and F. Calmon. Eγ-mixing time.
    Accepted for presentation at the IEEE International Symposium on Information Theory (ISIT), 2024.
     
  3. L. Monteiro, R. Cruz, F. Calmon and M. Diaz. On the inevitability of the Rashomon effect.
    IEEE International Symposium on Information Theory (ISIT), 2023. DOI:10.1109/ISIT54713.2023.10206657
     
  4. M. Diaz, P. Kairouz, J. Liao and L. Sankar. Neural network-based estimation of the MMSE.
    IEEE International Symposium on Information Theory (ISIT), 2021. DOI:10.1109/ISIT45174.2021.9518063
     
  5. H. Wang, H. Hsu, M. Diaz and F. Calmon. The impact of split classifiers on group fairness.
    IEEE International Symposium on Information Theory (ISIT), 2021. DOI: 10.1109/ISIT45174.2021.9517723
     
  6. S. Asoodeh, M. Diaz and F. Calmon. Privacy amplification of iterative algorithms via contraction coefficients.
    IEEE International Symposium on Information Theory (ISIT), 2020. DOI:10.1109/ISIT44484.2020.9174133
     
  7. T. Sypherd, M. Diaz, L. Sankar and G. Dasarathy. On the α-loss landscape in the logistic model.
    IEEE International Symposium on Information Theory (ISIT), 2020. DOI:10.1109/ISIT44484.2020.9174356
     
  8. T. Sypherd, M. Diaz, L. Sankar and P. Kairouz. A tunable loss function for binary classification.
    IEEE International Symposium on Information Theory (ISIT), 2019. DOI:10.1109/ISIT.2019.8849796
     
  9. H. Wang, M. Diaz, JCS Santos Filho and F. Calmon. An information-theoretic view of generalization via Wasserstein distance.
    IEEE International Symposium on Information Theory (ISIT), 2019. DOI:10.1109/ISIT.2019.8849359
     
  10. H. Wang, M. Diaz, F. Calmon and L. Sankar. The utility cost of robust privacy guarantees.
    IEEE International Symposium on Information Theory (ISIT), 2018. DOI:10.1109/ISIT.2018.8437735
     
  11. S. Asoodeh, M. Diaz, F. Alajaji and T. Linder. Privacy-aware guessing efficiency.
    IEEE International Symposium on Information Theory (ISIT), 2017. DOI:10.1109/ISIT.2017.8006629
     
  12. M. Diaz. On the symmetries and the capacity achieving input covariance matrices of multiantenna channels.
    IEEE International Symposium on Information Theory (ISIT), 2016. DOI:10.1109/ISIT.2016.7541464

Tutorials

  1. S. Asoodeh, F. Calmon, M. Diaz and H. Jeong. Information-Theoretic Tools for Responsible Machine Learning.
    IEEE International Symposium on Information Theory (ISIT), 2022. [website]
    –> Updated Slides (10/22) [pdf]

Teaching

Students

Miscellaneous Academic Activities

Last update: May 2024