Publications [Google Scholar]

* equal contribution authors. alphabetical authorship.

Papers

• A. Ganesh, M. Haghifam, T. Steinke, A. Thakurta ‘‘ Faster Differentially Private Convex Optimization via Second-Order Methods’’, preprint, 2023.

• A. Ganesh, M. Haghifam, M. Nasr, S. Oh, T. Steinke, O. Thakkar, A. Thakurta, L. Wang ‘‘ Why Is Public Pretraining Necessary for Private Model Training?’’, preprint, 2023 [paper]

• M. Haghifam*, B. Rodriguez-Galvez*, R. Thobaben, M. Skoglund, D. M. Roy, G. K. Dziugaite ‘‘Limitations of Information-Theoretic Generalization Bounds for Gradient Descent Methods in Stochastic Convex Optimization’’, International Conference on Algorithmic Learning Theory 34 (ALT), 2023 [paper].

• M. Haghifam, S. Moran, D. M. Roy, G. K. Dziugaite, ‘‘Understanding Generalization via Leave-One-Out Conditional Mutual Information’’, International Symposium on Information Theory 2022 [paper].

• M.Haghifam*, M. N. Krishnan*, A. Khisti, X. Zhu, W. Dan, J. Apostolopoulos, ‘‘On Streaming Codes With Unequal Error Protection’’, IEEE Journal on Selected Areas in Information Theory (Volume: 2, Issue: 4, Dec. 2021) [paper].

• M. Haghifam, V. Y. F. Tan, A. Khisti, ‘‘Sequential Classification with Empirically Observed Statistics’’, IEEE Transactions on Information Theory (Volume: 67, Issue: 5, May 2021) [paper].

• M. Haghifam, G. K. Dziugaite, S. Moran, D. M. Roy, ‘‘Towards a Unified Information–Theoretic Framework for Generalization’’, Advances in Neural Information Processing Systems 35 (NeurIPS), 2021 (Spotlight, <3% of submissions) [paper].
  (Also Presented as a workshop paper at ICML-21 Workshop on Information-Theoretic Methods for Rigorous, Responsible, and Reliable Machine Learning.)

• G. Neu, G. K. Dziugaite, M. Haghifam, D. M. Roy , ‘‘Information-Theoretic Generalization Bounds for Stochastic Gradient Descent’’, Annual Conference on Learning Theory 34 (COLT), 2021 [paper].

• M. Haghifam, J. Negrea, A. Khisti, D. M. Roy , G. K. Dziugaite, ‘‘Sharpened Generalization Bounds based on Conditional Mutual Information and an Application to Noisy, Iterative Algorithms’’, Advances in Neural Information Processing Systems 34 (NeurIPS), 2020 [paper].

• J. Negrea*, M. Haghifam*, G. K. Dziugaite, A. Khisti, D. M. Roy, ‘‘Information-Theoretic Generalization Bounds for SGLD via Data-Dependent Estimates’’, Advances in Neural Information Processing Systems 33 (NeurIPS), 2019 [paper].
  (Also Presented as a workshop paper at ICML-19 Workshop in Understanding and Improving Generalization in Deep Learning.)

• M. Haghifam, M. Robat Mili, B. Makki, M. Nasiri-Kenari, T. Svensson, ‘‘Joint Sum Rate And Error Probability Optimization: Finite Blocklength Analysis’’, IEEE Wireless Communications Letters (Volume: 6, Issue: 6, Dec. 2017) [paper].

Workshop Papers

• M. Haghifam, V. Y. F. Tan, and A. Khisti, ‘‘Sequential Classification with Empirically Observed Statistics’’, IEEE Information Theory Workshop 2019, Visby, Gotland, Sweden [paper].

• M. Haghifam, A. Badr, A. Khisti, X. Zhu, W. Dan and J. Apostolopoulos, ‘‘Streaming Codes with Unequal Error Protection against Burst Losses’’, The 29th Biennial Symposium on Communications (BSC 2018) [paper].

• M. Haghifam, B. Makki, M. Nasiri-Kenari, T. Svensson, ‘‘On joint information and energy transfer in relay networks with an imperfect power amplifier’’, 27th Annual IEEE International Symposium on Personal, Indoor, and Mobile Radio Communications (PIMRC), Valencia, Spain, 2016 [paper].

• M. Haghifam, M. R. Haghifam, B. Safari Chabook, ‘‘State estimation in electric distribution networks in presence of distributed generation using the PMUs’’, CIRED 2012, Lisbon, Portugal [paper]