I am a postdoc at the Harvard School of Engineering and Applied Science (SEAS), where I work with Flavio Calmon. Prior to that, I spent a year and half in Knowledge Lab and Institute of Genomics and Systems Biology at The University of Chicago as a postdoc. I received my PhD in Applied Math from Queen's University, where I was advised by Fady Alajaji and Tamas Linder.

My research combines information theory, machine learning, optimization, and mechanism design. I develop methods to provably promote the privacy and fairness of users in large-scale algorithms. 

News:

  • [Feb. 2021] Our new work on graph nural networks for soft semi-supervised learning has been accepted in Pacific-Asia KDD. [supplementary] [code]
  • [Feb. 2021] Check out our new work that shows local differential privacy is equivalent to the contration of HS divergence and gives tighter bounds for the minimax and Bayesian risks in private estimation problems  
  • [Feb. 2021] If you are interested in differential privacy and federated learning, have a look at our new paper here  
  • [Jan. 2021] I gave a talk at Harvard's Privacy Tools on the equivalency of local differential privacy and the contration of HS divergence [slides
  • [Jan. 2021] Our work was accepted in IEEE Journal on Selected Areas in Information Theory [Special Issue on Privacy and Security of Information Systems]  [IEEE's version]
  • [Jan. 2021] I gave a talk at Google on some new information-theoretic techniques for privacy analysis of iterative algorithms [slides
  • [Dec. 2020] Check out our new work on the application of strong data processing inequality to privacy analysis of online learning algorithms
  • [Oct. 2020] I am delighted to be selected as a top 10% high-scoring reviewer of NeurIPS 2020  
  • [Oct. 2020] Check out our new work on the generalization of information bottleneck and privacy funnel
  • [Aug. 2020] I will be in the program committe of AAAI 2021
  • [Aug. 2020] Check out our new work on the relationship between approximate DP, Renyi DP and hypothesis test DP
  • [July 2020]  I'll give a talk in ICML-FL workshop about our ongoing work on differentially private federated learning through an information theoretic lens   

 

<embed>
Copy and paste this code to your website.
Copy and paste this code to your website.