Decentralized Causal Discovery using Judo Calculus
Sridhar Mahadevan · 2025
Read Decentralized Causal Discovery using Judo Calculus by Sridhar Mahadevan (2025) 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-2510-23942.txt.
Description
Sridhar Mahadevan describes judo calculus, a framework for causal discovery that treats causal claims as locally true rather than true everywhere. In real domains — biology, medicine, social science — effects depend on regime: age, country, dose, genotype, or lab protocol. Classical methods flatten that context away.
Judo calculus keeps the context, using sheaf theory and a logic where truth is local. A modal operator picks the family of regimes that matter, and a claim counts as proven when it holds across that cover in a stable, constructive way. The framework plugs into standard score-based, constraint-based, and gradient-based discovery tools rather than replacing them. Tests range from synthetic data to real sets in biology and economics. The author reports faster runs from the decentralized design and better accuracy than classical methods — causal knowledge that knows where it holds.
Catalog record
| CREATOR | Sridhar Mahadevan |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 170 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-23942