Neuro-Symbolic Learning for Predictive Process Monitoring via Two-Stage Logic Tensor Networks with Rule Pruning
Fabrizio De Santis ✉, Gyunam Park, Francesco Zanichelli · 2026
Read Neuro-Symbolic Learning for Predictive Process Monitoring via Two-Stage Logic Tensor Networks with Rule Pruning by Fabrizio De Santis ✉, Gyunam Park, Francesco Zanichelli (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-2603-26944.txt.
Description
De Santis and co-authors add domain rules to predictive process monitoring through Logic Tensor Networks. Data-only models miss the sequential constraints that govern real event logs. Their method writes control-flow, temporal, and payload knowledge as logical formulas. A two-stage training scheme protects accuracy while the rules are learned. First, weighted axiom loss favors the data. Then rule pruning keeps only axioms that stay consistent and helpful. On four real event logs, the knowledge clearly improved predictions. Without the two stages, rules could badly hurt accuracy.
Catalog record
| CREATOR | Fabrizio De Santis ✉, Gyunam Park, Francesco Zanichelli |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 31330 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.26944