We are pleased to announce that Armeen Taeb, Assistant Professor of Statistics, has received grant funding support from the Royalty Research Fund (RRF) for his proposal “False discovery control for causal structure learning”.
Causal structure learning is the task of learning causal relationships from data. While causal structure learning has been studied extensively in the literature, we lack a systematic approach to measure and control false discoveries. In particular, due to global constraints imposed by acyclicity and equivalence class characterizations, the standard paradigm of formulating and testing a collection of binary hypothesis tests is not appropriate in this setting.
In this proposal, as the Principal Investigator (PI), Taeb aims to address previous shortcomings via an order-theoretic framework for model selection. Specifically, collections of causal models can be viewed as a partially ordered set (poset); the poset structure enables a hierarchical organization of the set of models and a natural definition of model complexity. Building on this observation, this proposal provides a formalism for measuring true and false discoveries in causal learning and develops algorithms to control false discoveries.
Congratulations to Armeen Taeb on his success!
Taeb receives Royalty Research Fund (RRF) Award
