Accompanying Research Paper Pending Publication
These are the results from running experiments to compare Symbolic Learning and Statistical Learning. We used Answer Set Grammars and solved these with ILASP for symbolic learning and a Random Forest, General Linear Model and a Fully Connected Feed-Forward Network for statistical learning.
The results were obtained from a machine with the following set of specifications:
Hardware
- Intel Xeon CPU E7-8870 @ 2.40GHz (80 cores)
- 576GB RAM DDR3 1333
Operating System
- Fedora 29 (x86_64)
- Kernel 5.1.11-200
Software
- python 3.7.3
- keras 2.4.2
- tensorflow 1.14.0 (cpu)
- h2o 3.24.0.5
- pandas 0.23.4
- scikit-learn 0.19.1
- ILASP version 3.4 -- beta 20/04/2019