Tag
This paper introduces ArborEnum, the first algorithm to exactly enumerate decision-tree Rashomon sets over continuous features without binarization, along with relaxed and anytime approximations that achieve orders-of-magnitude speedups while preserving near-perfect recall.
PRAXIS is a new algorithm that efficiently approximates the Rashomon set of near-optimal decision trees, achieving orders of magnitude improvement in runtime and memory while maintaining near-perfect recall.
Introduces horizon-constrained Rashomon sets to characterize how model multiplicity evolves in chaotic systems. The framework proves exponential contraction of predictive equivalence and develops decision-aligned algorithms that improve decision quality by 18-34%.