Research Project Title:
Learning Bayesian Networks with Observational and Interventional Data
abstract:In genomics, a question we wish to answer is: how do different genes interact? To answer this question, we seek to learn gene regulatory networks – causal directed acyclic graph (DAG) representations of genes, regulators, and the relationships between them. Learning such models requires a metric for measuring how well a model represents a given set of data and variables. For Gaussian Bayesian networks, where the conditional probability distributions are Gaussian, this metric has been computed when the dataset consists of observational data. In this SuperUROP, we wish to compute such a metric when we have access to both observational and interventional data in order to learn more representative models with more information.
Through this SuperUROP, I hope to learn more about causal inference and its applications in computational biology. I have taken various advanced math, statistics, and computer science courses, and I am eager to apply my past experiences to my project!