A Theoretical Comparison of No-U-Turn Sampler Variants: Necessary and Sufficient Convergence Conditions and Mixing Time Analysis under Gaussian Targets
Samuel Gruffaz, Kyurae Kim, Fares Guehtar, Hadrien Duval-decaix, Pacôme Trautmann · 2026
Read A Theoretical Comparison of No-U-Turn Sampler Variants: Necessary and Sufficient Convergence Conditions and Mixing Time Analysis under Gaussian Targets by Samuel Gruffaz, Kyurae Kim, Fares Guehtar, Hadrien Duval-decaix, Pacôme Trautmann (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-2603-18640.txt.
Description
Gruffaz and co-authors compare the two main variants of the No-U-Turn sampler. NUTS-mul picks indices by multinomial sampling, while NUTS-BPS uses biased progressive sampling. The team derives first necessary conditions for geometric ergodicity in both. They also prove sufficient conditions for NUTS-mul. For NUTS-BPS, they give a first mixing time result on Gaussian targets. Both variants mix in time that grows like the fourth root of dimension. The constants favor NUTS-BPS, though the broad behavior matches closely.
Catalog record
| CREATOR | Samuel Gruffaz, Kyurae Kim, Fares Guehtar, Hadrien Duval-decaix, Pacôme Trautmann |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 128495 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.18640