2022] I have been awarded the Scenario
Modeling Consortium Fellowship of $150,000 to support my contributions
in the projections of COVID-19, influenza, and future pathogens.
2022] I have been awarded an NSF RAPID grant
of $199,167 titled: “Data-driven Understanding of Imperfect Protection
for Long-term COVID-19 Projections”. This project will use models and machine
learning on vaccine breakthrough data and reinfection data to understand the
dynamics driving the reduction in COVID-19 immunity.
2021] My work with Scenario
Modeling Hub was used by ACIP
as part of the evidence to recommend vaccines for 5-11 years old children.
2021] I have been awarded an NSF RAPID grant
of $186,835 titled: "Fast COVID-19 Scenario Projections in Presence
of Vaccines and Competing Variants". This work will support my
contribution to the US/CDC
Scenario Modeling Hub in projecting the long-term impact of variant
dynamics, vaccines, waning immunity, etc.
is a word cloud of the titles of my paper. For the full list of my publications,
please see my Google
research interests include Machine Learning and Graph Algorithms applied
to epidemics, social good, and social networks. In the past, I have worked on
information diffusion, parallel computing, FPGA acceleration, and smartgrids. If you are a student interested in working
with me, please send me your resume and a half-page research proposal on what
you would like to pursue.
current work includes the following:
Forecasting the trajectory of an epidemic in the presence of changing policies,
under-reporting, competing variants, waning immunity, and vaccinations.
Year-long effort of many teams around the world with the CDC has shown that
there is high disagreement between different forecasting methodologies.
Identifying the peak and dynamics during multiple competing variants is
difficult. ML methods in the application often overfit. Research challenges: (i)
Learning model parameters without overfitting! (ii) Developing sophisticated
and explainable ensemble models to utilize multiple forecasts (iii) Developing
models of temporal dynamics of competing variants and parameter estimation from
Machine learning, epidemiological models, probability, GUI design
Selected Publications: KDD 2020
(not peer-reviewed), GitHub
Contributing to the CDC,
US forecast Hub, Germany+Poland
Deep GNNs on Large Graphs: Problem: Training and inference on deep GNNs on large graphs are difficult due to
computational complexity and lack of accuracy improvements with deeper layers.
Subgraph-based methods to address training on large graphs exist, but they do
not apply during inference, making inference the bottleneck. Such methods also
do not address poor accuracy for deep networks due to “oversmoothing”.
Research challenges: (i) Develop subgraph-based
schemes that apply to training and inference. (ii) Identify sampling
probabilities that eliminates oversmoothing
(converging embedded nodes to the same point in space) and maximizes
separation. (iii) Pruning weights to reduce computations during inference.
2021, VLDB 2021, ICLR 2020, IPDPS
Areas I wish to pursue in the future
1. Addressing homelessness SNAM
fairness in machine learning and algorithms.
3. In general,
I am happy to collaborate on machine learning, algorithms, and network science
applied to real-world problems.
In Spring 2022,
I am co-teaching “EE 638: Applications of Machine Learning for Medical Data”
with Prof. Cauligi Raghavendra.
In Fall 2021, I
taught “EE155L: Introduction to Computer Programming for Electrical Engineers”
with Prof. Sandeep Gupta.