This repository contains the experimental code and data for the paper: Learning Logic Programs by Discovering Higher-Order Abstractions, Céline Hocquette, Sebastijan Dumančić, and Andrew Cropper, accepted at IJCAI24.

Requirements

SWI-Prolog

Clingo 5.5.0

pyswip

ortools

[minikanren]

Experimental results

The results are in the folder /results/{condition}/results/{run_id}.

• The training tasks (examples, bias, bk, and induced programs) are in the folder /results/{condition}/results/{run_id}/{batch}/train/{task_name}
• The testing tasks (examples, bias, bk, and induced programs) are in the folder /results/{condition}/results/{run_id}/{batch}/test/{task_name}

Usage

Hopper is in the folder hopper.

You can reproduce the plots with the command: python show_results.py

You can run the learning / training again with the command: python stevie_experiment.py. It will generate new training and testing tasks.