PlatoLTL: Learning to Generalize Across Symbols in LTL Instructions for Multi-Task RL
Jacques Cloete, Mathias Jackermeier, Ioannis Havoutis, Alessandro Abate · 2026
Read PlatoLTL: Learning to Generalize Across Symbols in LTL Instructions for Multi-Task RL by Jacques Cloete, Mathias Jackermeier, Ioannis Havoutis, Alessandro Abate (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-2601-22891.txt.
Description
Jacques Cloete and three co-authors teach agents to follow task rules written in a formal logic. The logic, linear temporal logic, states step-by-step duties over time, like doing one thing before another. Agents must handle new tasks built from symbols they have not seen in training.
Their system, PlatoLTL, aims to generalize across those symbols rather than memorize them. An agent that reads rules flexibly could take new instructions without fresh training each time. That skill matters for robots and software agents facing changing jobs. Generalizing across symbols is the crux: the agent has to grasp the rule structure, not memorize which symbol stood where before.
Catalog record
| CREATOR | Jacques Cloete, Mathias Jackermeier, Ioannis Havoutis, Alessandro Abate |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 85 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-22891