Recent updates
May 25, 2021 Here is the hackathon that I designed based on the work of Zoabi et al. to predict COVID infection using symptoms. It was great to see diverse solutions that participants used to tackle the problem. I learned a lot. Thank you! Hackathon | Tutorial
May 20, 2021 I am joining CambridgeSpark team as a part-time faculty. To kick things off, I designed a hackathon for the new batch of L7 learners. This hackathon is based on the work of Zoabi et al. to predict COVID infection using symptoms. It exposes the participants to (a) reading applied machine learning papers, (b) problems encountered while dealing with imbalanced dataset, (c) building a machine learning application end-to-end.
Apr 8, 2021 I released a series of blog posts on contact tracing to explain the work that we have been doing since the onset of COVID-19.
Jan 12, 2021 Our work on Proactive Contact Tracing has been accepted for a spotlight talk at ICLR, 2020. Arxiv: link
Oct 23, 2020 We have uploaded the details of Proactive Contact Tracing, our proposed framework for contact tracing designed to predict individual infectiousness. The results are based on simulations on COVI-AgentSim.
Oct 20, 2020 We have uploaded the details of COVI-AgentSim on arXiv here. COVI-AgentSim simulates Covid-19 outbreak and estiamtes the impacts of different contact tracing methods, specifically those that leverage COVI protocol.
Sep 25, 2020 Our paper on Hybrid Models for Learning to Branch is accepted at NeurIPS, 2020
Jun 20, 2020 We have uploaded a white paper on COVI, our proposed peer-to-peer communication protocol that enables better contact tracing. More on the methods in the following months.
Apr 1, 2020 Our work on deep metric learning baselines and training protocols is accepted at ICML, 2020. Read more about it here. This work was led by the mastermind Karsten Roth.
Jun 21, 2019 I am starting my research internship at Mila under the supervision of Prof. Andrea Lodi and Prof. Yoshua Bengio.
Sep 25, 2017 I am starting my Ph.D. under the supervision of Prof. M. Pawan Kumar at the University of Oxford, funded by The Alan Turing Institute.