Hafiz Asif

I am a post-doctoral associate at the institute of Data Science, Learning, and Applications at Rutgers. Here, I'm developing privacy-protecting approaches and tools to analyze spatiotemporal data streams, exploring their application in tracking infectious diseases, e.g.,  Covid-19 (CovidNearby.org), and developing privacy- and fairness-preserving models and algorithms to analyze healthcare and biomedical data.

Before that, I was a fellow at INRIA (the french national institute of research) Paris, France, but my term was cut short due to the pandemic. I received my Ph.D. from Rutgers University under the supervision of Jaideep Vaidya and Periklis Papakonstantinou. 

​

The overarching goal of my research is to develop socially-responsible AI and data-driven approaches to accelerate equitable biomedical research, enable joint analysis of federated sensitive data, and build fair decision models via machine learning. I am a practitioner at heart with solid theoretical foundations. and my research is driven by the impactful real-world problems. However, I also do deep theoretical innovations, which in many cases, is essential for solving practical privacy and security problems in the modern systems. 

 

To this end, I explicate models of privacy and security, fairness and equity, and employ them to design SAFE approaches  (SAFE=Secure, Auditable, Fair, and Equitable) for data-driven solutions, e.g., learning a model to identify fraud or assess an individual's risk of contracting an infection. SAFE approaches promise to make segregate, siloed, and sensitive data available for research and analysis. Here are a few problems that I'm working on right now. 

​

• Privacy-protecting spatiotemporal data analysis

  • ​How to effectivley surveille an infectious desiase, e.g., COVID-19, via health-related and location data but without relinquishing perople's privacy?

  • How to design utility-maximizing differentially private mechanisms for analyzing temoral and spatial correlations in the data?

• SAFE (Secure, Auditable, Fair, and Equitable) data analytics

  • How to leverage big data that is segregated across locked silos to carryout biomedical research via machine learning while assuring security and privacy of the data, auditibility of the computational processes, and fair and equitable applicablity of the results?​

​

Besides pursuing my research endevours, I love hiking, bouldering, traveling, and learning new magic tricks.