Neuro-Symbolic Hierarchical Rule Induction

Claire Glanois, Zhaohui Jiang, Xuening Feng, Paul Weng, Matthieu Zimmer, Dong Li, Wulong Liu, Jianye Hao
Proceedings of the 39th International Conference on Machine Learning, PMLR 162:7583-7615, 2022.

Abstract

We propose Neuro-Symbolic Hierarchical Rule Induction, an efficient interpretable neuro-symbolic model, to solve Inductive Logic Programming (ILP) problems. In this model, which is built from a pre-defined set of meta-rules organized in a hierarchical structure, first-order rules are invented by learning embeddings to match facts and body predicates of a meta-rule. To instantiate, we specifically design an expressive set of generic meta-rules, and demonstrate they generate a consequent fragment of Horn clauses. As a differentiable model, HRI can be trained both via supervised learning and reinforcement learning. To converge to interpretable rules, we inject a controlled noise to avoid local optima and employ an interpretability-regularization term. We empirically validate our model on various tasks (ILP, visual genome, reinforcement learning) against relevant state-of-the-art methods, including traditional ILP methods and neuro-symbolic models.

Cite this Paper

BibTeX
@InProceedings{pmlr-v162-glanois22a, title = {Neuro-Symbolic Hierarchical Rule Induction}, author = {Glanois, Claire and Jiang, Zhaohui and Feng, Xuening and Weng, Paul and Zimmer, Matthieu and Li, Dong and Liu, Wulong and Hao, Jianye}, booktitle = {Proceedings of the 39th International Conference on Machine Learning}, pages = {7583--7615}, year = {2022}, editor = {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan}, volume = {162}, series = {Proceedings of Machine Learning Research}, month = {17--23 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v162/glanois22a/glanois22a.pdf}, url = {https://proceedings.mlr.press/v162/glanois22a.html}, abstract = {We propose Neuro-Symbolic Hierarchical Rule Induction, an efficient interpretable neuro-symbolic model, to solve Inductive Logic Programming (ILP) problems. In this model, which is built from a pre-defined set of meta-rules organized in a hierarchical structure, first-order rules are invented by learning embeddings to match facts and body predicates of a meta-rule. To instantiate, we specifically design an expressive set of generic meta-rules, and demonstrate they generate a consequent fragment of Horn clauses. As a differentiable model, HRI can be trained both via supervised learning and reinforcement learning. To converge to interpretable rules, we inject a controlled noise to avoid local optima and employ an interpretability-regularization term. We empirically validate our model on various tasks (ILP, visual genome, reinforcement learning) against relevant state-of-the-art methods, including traditional ILP methods and neuro-symbolic models.} }
Endnote
%0 Conference Paper %T Neuro-Symbolic Hierarchical Rule Induction %A Claire Glanois %A Zhaohui Jiang %A Xuening Feng %A Paul Weng %A Matthieu Zimmer %A Dong Li %A Wulong Liu %A Jianye Hao %B Proceedings of the 39th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2022 %E Kamalika Chaudhuri %E Stefanie Jegelka %E Le Song %E Csaba Szepesvari %E Gang Niu %E Sivan Sabato %F pmlr-v162-glanois22a %I PMLR %P 7583--7615 %U https://proceedings.mlr.press/v162/glanois22a.html %V 162 %X We propose Neuro-Symbolic Hierarchical Rule Induction, an efficient interpretable neuro-symbolic model, to solve Inductive Logic Programming (ILP) problems. In this model, which is built from a pre-defined set of meta-rules organized in a hierarchical structure, first-order rules are invented by learning embeddings to match facts and body predicates of a meta-rule. To instantiate, we specifically design an expressive set of generic meta-rules, and demonstrate they generate a consequent fragment of Horn clauses. As a differentiable model, HRI can be trained both via supervised learning and reinforcement learning. To converge to interpretable rules, we inject a controlled noise to avoid local optima and employ an interpretability-regularization term. We empirically validate our model on various tasks (ILP, visual genome, reinforcement learning) against relevant state-of-the-art methods, including traditional ILP methods and neuro-symbolic models.
APA
Glanois, C., Jiang, Z., Feng, X., Weng, P., Zimmer, M., Li, D., Liu, W. & Hao, J.. (2022). Neuro-Symbolic Hierarchical Rule Induction. Proceedings of the 39th International Conference on Machine Learning, in Proceedings of Machine Learning Research 162:7583-7615 Available from https://proceedings.mlr.press/v162/glanois22a.html.

Related Material