Saved in:
| Main Author: | |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2506.11721 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908407265492992 |
|---|---|
| author | Chen, Dillon Z. |
| author_facet | Chen, Dillon Z. |
| contents | Relational Graph Neural Networks (R-GNNs) are a GNN-based approach for learning value functions that can generalise to unseen problems from a given planning domain. R-GNNs were theoretically motivated by the well known connection between the expressive power of GNNs and $C_2$, first-order logic with two variables and counting. In the context of planning, $C_2$ features refer to the set of formulae in $C_2$ with relations defined by the unary and binary predicates of a planning domain. Some planning domains exhibit optimal value functions that can be decomposed as arithmetic expressions of $C_2$ features. We show that, contrary to empirical results, R-GNNs cannot learn value functions defined by $C_2$ features. We also identify prior GNN architectures for planning that may better learn value functions defined by $C_2$ features. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_11721 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Relational GNNs Cannot Learn $C_2$ Features for Planning Chen, Dillon Z. Artificial Intelligence Machine Learning Relational Graph Neural Networks (R-GNNs) are a GNN-based approach for learning value functions that can generalise to unseen problems from a given planning domain. R-GNNs were theoretically motivated by the well known connection between the expressive power of GNNs and $C_2$, first-order logic with two variables and counting. In the context of planning, $C_2$ features refer to the set of formulae in $C_2$ with relations defined by the unary and binary predicates of a planning domain. Some planning domains exhibit optimal value functions that can be decomposed as arithmetic expressions of $C_2$ features. We show that, contrary to empirical results, R-GNNs cannot learn value functions defined by $C_2$ features. We also identify prior GNN architectures for planning that may better learn value functions defined by $C_2$ features. |
| title | Relational GNNs Cannot Learn $C_2$ Features for Planning |
| topic | Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2506.11721 |