Showing 2024Show all
3 papers · 1 filter
cs.LG2024
The Star Geometry of Critic-Based Regularizer Learning
Oscar Leong, Eliza O'Reilly, Yong Sheng Soh
Variational regularization is a classical technique to solve statistical inference tasks and inverse problems, with modern data-driven approaches parameterizing regularizers via de…
math.ST2024
TrIM: Transformed Iterative Mondrian Forests for Gradient-based Dimension Reduction and High-Dimensional Regression
Ricardo Baptista, Eliza O'Reilly, Yangxinyu Xie
We propose a computationally efficient algorithm for gradient-based linear dimension reduction and high-dimensional regression. The algorithm initially computes a Mondrian forest a…
math.ST2024
Statistical Advantages of Oblique Randomized Decision Trees and Forests
Eliza O'Reilly
This work studies the statistical implications of using features comprised of general linear combinations of covariates to partition the data in randomized decision tree and forest…