Learn to Explain Efficiently via Neural Logic Inductive Learning
Abstract
The capability of making interpretable and self-explanatory decisions is essential for developing responsible machine learning systems. In this work, we study the learning to explain the problem in the scope of inductive logic programming (ILP). We propose Neural Logic Inductive Learning (NLIL), an efficient differentiable ILP framework that learns first-order logic rules that can explain the patterns in the data. In experiments, compared with the state-of-the-art models, we find NLIL is able to search for rules that are x10 times longer while remaining x3 times faster. We also show that NLIL can scale to large image datasets, i.e. Visual Genome, with 1M entities.
Cite
Text
Yang and Song. "Learn to Explain Efficiently via Neural Logic Inductive Learning." International Conference on Learning Representations, 2020.Markdown
[Yang and Song. "Learn to Explain Efficiently via Neural Logic Inductive Learning." International Conference on Learning Representations, 2020.](https://mlanthology.org/iclr/2020/yang2020iclr-learn/)BibTeX
@inproceedings{yang2020iclr-learn,
title = {{Learn to Explain Efficiently via Neural Logic Inductive Learning}},
author = {Yang, Yuan and Song, Le},
booktitle = {International Conference on Learning Representations},
year = {2020},
url = {https://mlanthology.org/iclr/2020/yang2020iclr-learn/}
}