ElliCE: Efficient and Provably Robust Algorithmic Recourse via the Rashomon Sets
Bohdan Turbal, Iryna Voitsitska, Lesia Semenova · 2026
Read ElliCE: Efficient and Provably Robust Algorithmic Recourse via the Rashomon Sets by Bohdan Turbal, Iryna Voitsitska, Lesia Semenova (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-2602-07674.txt.
Description
When a model denies someone a loan or a job, algorithmic recourse tells them what to change for a better outcome. Standard counterfactuals can fail if another near-optimal model would decide otherwise. Bohdan Turbal and two co-authors present ElliCE, a method for recourse that stays valid across the Rashomon set of near-optimal models.
Optimizing counterfactuals over an ellipsoidal approximation of that set yields explanations provably valid across the ellipsoid. Guarantees on uniqueness, stability, and match with key feature directions, backed by tests. Responsible AI teams get recourse that survives model multiplicity instead of depending on one arbitrary model among many equally good ones.
Catalog record
| CREATOR | Bohdan Turbal, Iryna Voitsitska, Lesia Semenova |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 117 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-07674