Evaluating Logical Generalization in Graph Neural Networks

Abstract

Recent research has highlighted the role of relational inductive biases in building learning agents that can generalize and reason in a compositional manner. However, while relational learning algorithms such as graph neural networks (GNNs) show promise, we do not understand how effectively these approaches can adapt to new tasks. In this work, we study the task of logical generalization using GNNs by designing a benchmark suite grounded in first-order logic. Our benchmark suite, GraphLog, requires that learning algorithms perform rule induction in different synthetic logics, represented as knowledge graphs. GraphLog consists of relation prediction tasks on 57 distinct logical domains. We use GraphLog to evaluate GNNs in three different setups: single-task supervised learning, multi-task pretraining, and continual learning. Unlike previous benchmarks, our approach allows us to precisely control the logical relationship between the different tasks. We find that the ability for models to generalize and adapt is strongly determined by the diversity of the logical rules they encounter during training, and our results highlight new challenges for the design of GNN models. We publicly release the dataset and code used to generate and interact with the dataset at https://www.cs.mcgill.ca/~ksinha4/graphlog.

References in corpus (12)

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Semi-Supervised Classification with Graph Convolutional Networks
• Neural Message Passing for Quantum Chemistry
• A continual learning survey: Defying forgetting in classification tasks
• Theoretical Models of Learning to Learn
• Interaction Networks for Learning about Objects, Relations and Physics
• Differentiable Learning of Logical Rules for Knowledge Base Reasoning
• Graph Neural Tangent Kernel: Fusing Graph Neural Networks with Graph Kernels
• DRUM: End-To-End Differentiable Rule Mining On Knowledge Graphs
• Embedding Logical Queries on Knowledge Graphs
• Are Powerful Graph Neural Nets Necessary? A Dissection on Graph Classification
• Enhancing the Transformer with Explicit Relational Encoding for Math Problem Solving

Cited by in corpus (6)

• Open Graph Benchmark: Datasets for Machine Learning on Graphs
• Machine Learning on Graphs: A Model and Comprehensive Taxonomy
• Relational Graph Convolutional Networks: A Closer Look
• From Local Structures to Size Generalization in Graph Neural Networks
• Knowledge Hypergraph Embedding Meets Relational Algebra
• The Role of Isomorphism Classes in Multi-Relational Datasets