My research focuses on designing domain adaptation & transfer learning models for high-dimensional spatio-temporal datasets.
The current transfer learning methodologies do not account for non-i.i.d characteristics in the data. Since, spatio-temporal data often consists of
such spurious characterisitcs, the objective of my research is to understand latent patterns in the data, improve the robustness of transfer learning
techniques and design effective (data-efficient) modeling techniques for the (complex) data. A problem that I am keenly interested in is designing transfer learning
models for the prediction of air pollution in countries/communities that require a cost effective solution.