Ensemble-Conditional Gaussian Processes (Ens-CGP): Representation, Geometry, and Inference
Sai Ravela, Jae Deok Kim, Kenneth Gee, Xingjian Yan, Samson Mercier, Lubna Albarghouty, Anamitra Saha · 2026
Read Ensemble-Conditional Gaussian Processes (Ens-CGP): Representation, Geometry, and Inference by Sai Ravela, Jae Deok Kim, Kenneth Gee, Xingjian Yan, Samson Mercier, Lubna Albarghouty, Anamitra Saha (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-2602-13871.txt.
Description
Sai Ravela and six co-authors set out a 2026 arXiv paper on a unified view of ensemble inference. At its center sits the conditional Gaussian law, the full posterior for a Gaussian prior with linear observations. Kalman filtering, they stress, is just one way to compute that law under dynamics, not the law itself.
Their synthesis, called Ens-CGP, builds the prior from ensemble moments, which may be low-rank, and then conditions exactly. By splitting representation from computation, the framework links Kalman filters, ensemble variants, and iterative schemes to one object. It also ties the same law to MAP estimation, kernel-style regression, and classic regularization, clarifying how probabilistic and variational views fit together.
Catalog record
| CREATOR | Sai Ravela, Jae Deok Kim, Kenneth Gee, Xingjian Yan, Samson Mercier, Lubna Albarghouty, Anamitra Saha |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 41 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-13871