Research Project Title:
Inferring Structure in Gaussian Graphical Models from Noisy Observations
abstract:In this project, we develop an algorithm targeted at inferring the latent structure of directed acyclic Gaussian graphical models from observations through a noise function. This algorithm has direct applications in computational biology, where inferring the structure of gene regulatory networks — the causal relationship between the genes expressed in each cell — is of great importance. We provide guarantees on the consistency of our algorithm, and evaluate its performance in controlled simulations. Moreover, we display its operation on a single-cell RNA-seq data-set.
“I’d like to participate in a research project where I hold a great deal of the responsibility, and I think going through this program in this lab will provide the necessary mentorship for me to be successful in doing that.”