53 citations · 56 across the 8 of their papers we have counts for
8 papers
The -test: Increasing the Linear Model -test's Power Under Sparsity Without Sacrificing Validity
Danielle Paulson, Souhardya Sengupta, Lucas Janson
We introduce a new procedure for testing the significance of a set of regression coefficients in a Gaussian linear model with . Our method, the -test, provides the sam…
Chiseling: Powerful and Valid Subgroup Selection via Interactive Machine Learning
Nathan Cheng, Asher Spector, Lucas Janson
In regression and causal inference, controlled subgroup selection aims to identify, with inferential guarantees, a subgroup (defined as a subset of the covariate space) on which th…
The Relative Instability of Model Comparison with Cross-validation
Alexandre Bayle, Lucas Janson, Lester Mackey
Cross-validation (CV) is known to provide asymptotically exact tests and confidence intervals for model improvement but only when the model comparison is relatively stable. Surpris…
Semiparametric Inference for Partially Identifiable Data Fusion Estimands via Double Machine Learning
Yicong Jiang, Lucas Janson
Many statistical estimands of interest (e.g., in regression or causality) are functions of the joint distribution of multiple random variables. But in some applications, data is no…
Optimal ablation for interpretability
Maximilian Li, Lucas Janson
Interpretability studies often involve tracing the flow of information through machine learning models to identify specific model components that perform relevant computations for…
The -test: leveraging sparsity in the Gaussian linear model for improved inference
Souhardya Sengupta, Lucas Janson
We develop novel LASSO-based methods for coefficient testing and confidence interval construction in the Gaussian linear model with . Our methods' finite-sample validity is…
Total Variation Floodgate for Variable Importance Inference in Classification
Wenshuo Wang, Lucas Janson, Lihua Lei +1
Inferring variable importance is the key problem of many scientific studies, where researchers seek to learn the effect of a feature on the outcome in the presence of confo…
Effective Degrees of Freedom: A Flawed Metaphor
Lucas Janson, William Fithian, Trevor Hastie
To most applied statisticians, a fitting procedure's degrees of freedom is synonymous with its model complexity, or its capacity for overfitting to data. In particular, it is often…