Embracing Discrete Search: A Reasonable Approach to Causal Structure Learning

Researchers introduce FLOP, a new causal discovery algorithm for linear models that significantly reduces computation time through fast parent selection and Cholesky-based score updates. The algorithm achieves near-perfect accuracy in standard benchmarks and makes discrete search approaches viable for causal structure learning.

• →FLOP algorithm dramatically reduces run-times compared to prior causal discovery methods
• →The approach enables effective discrete search over graph structures for causal discovery
• →Algorithm demonstrates near-perfect recovery rates in standard benchmark settings
• →Results suggest discrete search methods deserve reconsideration in causal discovery research
• →Fast parent selection paired with iterative score updates makes comprehensive search feasible