Lifted Action Models Learning from Partial Traces

This repository contains the official code of the Offline Learning of Action Models (OffLAM) algorithm.

Installation

pip install offlam

Example usage

from offlam.algorithm import learn
model = learn('path/to/domain.pddl', ['path/to/trace0', 'path/to/trace1'])
print(model)

Custom domain learning

The OffLAM algorithm can be run for learning from traces with partially observable states, partially observable actions, and partially observable states and actions. For running OffLAM on a custom domain, you need to provide an input domain file 'path/to/domain.pddl' and a list of plan trace files ['path/to/trace0', 'path/to/trace1', etc.]. The input planning domain must contain the predicates, object types, and operator signatures, an example of (empty) input planning domain is Analysis/Benchmarks/testworld.pddl. Examples of input plan traces with partial states can be found in the directory offlam/Analysis/Input traces/testworld/partial_states, notice that OffLAM can learn a planning domain from plan traces of different environments (e.g. it is possible to learn a planning domain from small environments and exploit the learned domain in large environments).

Citations

@article{lamanna2024lifted,
  title={Lifted Action Models Learning from Partial Traces},
  author={Lamanna, Leonardo and Serafini, Luciano and Saetti, Alessandro and Gerevini, Alfonso and Traverso, Paolo},
  journal={Artificial Intelligence},
  volume={339},
  pages={104256},
  year={2025},
  publisher={Elsevier}
}

License

This project is licensed under the MIT License - see the LICENSE file for details.