Assistant Professor,

Computer Science & Engineering,

The Ohio State University

email: bassily.1@osu.edu

About me

My research interests span several areas such as privacy-preserving data analysis, machine learning, optimization, and information and coding theory. I am broadly interested in studying the tension/harmony between data analysis and machine learning on one hand and other notions of central importance to people and society such as privacy and security. In my work, I enjoy applying various tools from several areas such as learning theory, optimization, statistics, and information theory.

My research focuses on tackling current challenges in data analysis and machine learning especially those of direct impact on society. Much of my recent research effort has been devoted to developing practical algorithms with rigorous guarantees for privacy-preserving data analysis and machine learning. The goal of this area of research is to design highly accurate machine learning and data analysis algorithms that use private, personal data while providing rigorous guarantees of privacy for individuals whose data are collected; that is, to achieve the seemingly paradoxical goal of learning from private data without learning private data! Part of my research also addresses fundamental questions in machine learning and privacy.

Since Fall 2017, I have been an assistant professor in the Department of Computer Science and Engineering at The Ohio State University. Before joining OSU, I was a data-science postdoctoral fellow at University of California, San Diego. Prior to this, I was a postdoc in the Department of Computer Science and Engineering at The Pennsylvania State University. I completed my PhD in Electrical and Computer Engineering at University of Maryland, College Park, in 2012. I am a recipient of Google Faculty Research Award for year 2018.

News:

  • Sep 2020: Two papers accepted at NeurIPS 2020, and our paper on the Stability of SGD on Non-smooth Losses has been selected for spotlight presentation at the venue.

  • August 2020: A new paper on Learning from Mixtures of Private and Public Populations.

  • June 2020: A new paper on Stability of SGD on Nonsmooth Convex Losses.

  • June 2020: Our paper on Private Query Release Assisted by Public Data has been accepted to ICML 2020.

  • Nov 2019: Paper with my student Anupama Nandi on Privately Answering Classification Queries has been accepted to ALT 2020.

  • Sep 2019: Two papers accepted at NeurIPS 2019. Our paper on Private Stochastic Convex Optimization has been selected for spotlight presentation at NeurIPS 2019.

  • August 2019: Invited talk at the International Conference on Continuous Optimization (ICCOPT 2019), Berlin, Germany.

  • July 2019: Awarded NSF grants # 1908281 (NSF/AF - as PI) and # 1907715 (NSF/SHF & SaTC - as co-PI).

  • June 2019: Invited talk at Google Workshop on Federated Learning, Google Seattle.

  • April 2019: Invited talk at the Privacy and the Science of Data Analysis Workshop held at Simons Institute on Private Learning with Auxiliary Public Data.

  • March 2019: I received Google Faculty Research Award for year 2018 (under the "Privacy" category).

  • Jan 2019: During Spring 2019, I will be participating as a long-term visiting scholar in the semester-long Data Privacy Program at Simons Institute, Berkeley.

  • Dec 2018: I will be serving on the Program Committee for the 26th ACM Conference on Computer and Communications Security (CCS) 2019.

  • Dec 2018: My paper on Linear Queries with Local Differential Privacy has been accepted at AISTATS 2019.

  • Sep 2018: Our paper on Model-Agnostic Private Learning has been accepted for oral presentation at NEURIPS (formerly, NIPS) 2018.

  • July 2018: I am serving on the PC of TPDP 2018 (Workshop on Theory & Practice of Differential Privacy). Submission deadline (2-4 page abstract): July 27 (new date!). The workshop includes a wide range of topics revolving around data privacy. If you have results related to differential privacy, you are encouraged to submit your work by the deadline here!

  • June 2018: Our paper on mini-batch SGD for over-parameterized learning will be presented as a long talk at ICML!

  • April 2018: Participating in BIRS program on Mathematical Foundations of Data Privacy, Banff, CA [Video of my talk].

  • Jan 2018: Teaching a class on Privacy-Preserving Data Analysis and Differential Privacy this Spring!

  • Dec 2017: Our paper "Learners That Leak Little Information" has been accepted to ALT 2018!

  • Sep 2017: Our paper "Practical Locally Private Heavy Hitters" has been accepted at NIPS 2017!

  • Aug 2017: Teaching the Machine Learning Theory class CSE 5523 this Fall!

  • Aug 2017: Joined the CSE Department at OSU as Assistant Professor.

Theory Seminar:

Check out our theory seminar here.

Students:

  • Anupama Nandi (PhD): Summer 2018 - Present

  • Michael Menart (PhD): Fall 2019 -

Note to Prospective Students:

I am looking for bright, self-motivated PhD students to recruit. If you are interested in working with me on one or more of the topics in the link below, feel free to send me a brief email with your background and your resume (including any relevant coursework you have done). Also, taking one of my classes (e.g., CSE 5223: Machine learning theory this Fall) and doing very well is a plus. If you are not currently admitted to OSU, you are encouraged to apply. If that is the case, you are still welcome to send me an email, but, please keep in mind that admissions are decided by the department/university committees.

Some Research Topics: here