Exactly Computing do-Shapley Values
R. Teal Witter, Álvaro Parafita, Tomàs Garriga, Maximilian Muschalik, Fabian Fumagalli, Axel Brando, Lucas Rosenblatt · 2026
Read Exactly Computing do-Shapley Values by R. Teal Witter, Álvaro Parafita, Tomàs Garriga, Maximilian Muschalik, Fabian Fumagalli, Axel Brando, Lucas Rosenblatt (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-2602-07203.txt.
Description
Explaining a complex system often means asking how much each variable matters under interventions, and do-Shapley values answer that question game-theoretically. R. Teal Witter and six co-authors make those values exactly computable in far more cases than before.
The key is a reformulation in terms of the irreducible sets of the underlying structural causal model. The exact method runs in time linear in their number, which ranges from the number of variables up to exponential depending on graph structure. A companion estimator handles any query budget and converges as the budget grows. Causal analysts get both an exact tool and a graceful fallback for hard graphs.
Catalog record
| CREATOR | R. Teal Witter, Álvaro Parafita, Tomàs Garriga, Maximilian Muschalik, Fabian Fumagalli, Axel Brando, Lucas Rosenblatt |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 90 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-07203