Adel Javanmard
 |
Associate Professor Data Sciences and Operations Department (Statistics group) (by courtesy) Department of Computer Science |
Contact
Office: 300A Bridge Hall, University of Southern California, Los Angeles, CA 90089
Tel: (213)821-4193
Email: ajavanma(at)usc.edu
Here are links to my Google Scholar Profile, LinkedIn Profile and a short bio in the third-person.
About Me
I am an Associate Professor (with tenure) in the department of Data Sciences and Operations, Marshall School of Business at the University of Southern California (USC). I also hold a courtesy appointment at the Computer Science, USC Viterbi School of Engineering.
I also currently serve as an Associate Editor for Operations Research in the “Machine Learning and Data Science” department,
I am broadly interested in design and analysis of statistical methods for large-scale data, high-dimensional inference, network analysis, non-convex optimization and personalized decision-making. For more details, please see my publications.
Prior to joining USC, I was NSF CSoI postdoctoral fellow with worksite at Stanford University and UC Berkeley. I obtained my Ph.D. in Electrical Engineering from Stanford University advised by Andrea Montanari.
What's new?
January 2022: New paper on the curse of overparametrization in adversarial training.
November 2021: A new paper on PCR test. (No! nothing to do with COVID :) )
October 2021: New paper on theoretical understanding of adversarial training for latent models.
August 2021: Online debiasing accepted to the Journal of American Statistical Association!
August 2021: New paper on tree-based aggregation is out!
Controlling the False Split Rate in Tree-Based AggregationMay 2021: Fundamental Tradeoffs in Distributionally Adversarial Training is accepted at ICML 2021!
April 2021: I am honored to have been promoted to an Associate Professor with tenure!
February 2021: I'm honored to be selected as a 2021 Alfred P. Sloan Research Fellow in Mathematics.
December 2020: New paper on testing data privacy compliance of machine learning platforms.
Near-Optimal Model Discrimination with Non-DisclosureOctober 2020: New paper on the theory of adversarial training for classification is out!
Precise Statistical Analysis of Classification Accuracies for Adversarial TrainingSeptember 2020: Received an Adobe Data Science Faculty Research award. Thanks Adobe for your support!
August 2020: I am presenting the IMS Tweedie Invited Lecture at Bernoulli-IMS One World Symposium 2020 [Link]
Selected Publications
For the complete list of publications, check here
Precise Tradeoffs in Adversarial Training for Linear Regression
Adel Javanmard, Mahdi Soltanolkotabi, Hamed Hassani
Annual Conference on Learning Theory (COLT), 2020
Analysis of a Two-Layer Neural Network via Displacement Convexity
Adel Javanmard, Marco Mondelli, Andrea Montanari
Accepted for publication in Annals of Statistics, 2019
Dynamic Incentive-Aware Learning: Robust Pricing in Contextual Auctions
Negin Golrezaei, Adel Javanmard and Vahab Mirrokni
Accepted for publication in Operations Research, 2019.
(Preliminary version of this paper accepted to NeurIPS 2019.)
Theoretical insights into the optimization landscape of over-parameterized shallow neural networks
Mahdi Soltanolkotabi, Adel Javanmard and Jason D. Lee
in IEEE Transaction on Information Theory, 65(2), pages 742-769, 2018.
Phase Transitions in Semidefinite Relaxations [Website]
Adel Javanmard, Andrea Montanari and Federico Ricci-Tersenghi
In Proceedings of the National Academy of Sciences (PNAS), 113(16): E2218-E2223, 2016
Confidence Intervals and
Hypothesis Testing for High-Dimensional Regression [Website]
Adel Javanmard and Andrea Montanari
in Journal of Machine Learning Research, 15(1): 2869-2909, 2014.
Information-Theoretically Optimal Compressed Sensing via Spatial Coupling and Approximate Message Passing
David L. Donoho, Adel Javanmard, Andrea Montanari
IEEE Transaction on Information Theory, vol. 59, no. 11, pp 7434-7464, Nov 2013.