DARTS: Targeting Prognostic Covariates in Budget-Constrained Sequential Experiments
Kateryna Husar, Alexander Volfovsky · 2026
Read DARTS: Targeting Prognostic Covariates in Budget-Constrained Sequential Experiments by Kateryna Husar, Alexander Volfovsky (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-06608.txt.
Description
Kateryna Husar and one co-author design medical trials that must pay to measure patient covariates before treatment. Trials usually assume useful outcome-linked traits are known and free, but rich data taken before treatment are costly, forcing a trade-off between sharpness and budget. Their DARTS method treats covariate buying as a step-by-step choice inside the trial design itself.
A budgeted Thompson sampler learns across batches which traits predict outcomes best, and chosen traits drive fresh random splits plus regression to cut effect-estimate variance. Theory shows adaptive selection preserves randomization soundness and nominal coverage, with a near-optimal risk bound for the buying layer. In tests DARTS focuses spending on useful features and nears oracle gains.
Catalog record
| CREATOR | Kateryna Husar, Alexander Volfovsky |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 80 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.06608