Searching for actual causes: Approximate algorithms with adjustable precision
Samuel Reyd, Ada Diaconescu, Jean-Louis Dessalles · 2025
Read Searching for actual causes: Approximate algorithms with adjustable precision by Samuel Reyd, Ada Diaconescu, Jean-Louis Dessalles (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-2507-07857.txt.
Description
Samuel Reyd and co-authors searched for actual causes with algorithms in two thousand twenty-five. An actual cause explains what made this outcome happen. It differs from a general tendency across cases. Users of AI explanations expect that specific kind of answer. A classic theory by Halpern and Pearl formalized the idea. Finding such causes exactly is computationally hard.
The team offers an approximate polynomial-time algorithm. Its precision can be dialed up or down as needed. That flexibility suits real-time explanation tools. Past methods covered only narrow model classes. They also tended to return just the shortest cause. The new work widens both scope and output. It brings formal causation closer to daily AI use.
Catalog record
| CREATOR | Samuel Reyd, Ada Diaconescu, Jean-Louis Dessalles |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 84 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-07857