RELATE: A Schema-Agnostic Perceiver Encoder for Multimodal Relational Graphs

TL;DR

RELATE is a versatile, schema-agnostic encoder for heterogeneous relational graphs that enhances scalability and supports multi-dataset pretraining, achieving near state-of-the-art performance with fewer parameters.

Contribution

Introduces RELATE, a schema-agnostic, plug-and-play feature encoder using shared modality-specific encoders and cross-attention, enabling scalable and general-purpose GNNs for relational data.

Findings

Achieves within 3% of schema-specific encoders on benchmark tasks.

Reduces parameter count by up to 5x compared to existing methods.

Supports multi-dataset pretraining for relational graph models.

Abstract

Relational multi-table data is common in domains such as e-commerce, healthcare, and scientific research, and can be naturally represented as heterogeneous temporal graphs with multi-modal node attributes. Existing graph neural networks (GNNs) rely on schema-specific feature encoders, requiring separate modules for each node type and feature column, which hinders scalability and parameter sharing. We introduce RELATE (Relational Encoder for Latent Aggregation of Typed Entities), a schema-agnostic, plug-and-play feature encoder that can be used with any general purpose GNN. RELATE employs shared modality-specific encoders for categorical, numerical, textual, and temporal attributes, followed by a Perceiver-style cross-attention module that aggregates features into a fixed-size, permutation-invariant node representation. We evaluate RELATE on ReLGNN and HGT in the RelBench benchmark, where it achieves performance within 3% of schema-specific encoders while reducing parameter counts by up to 5x. This design supports varying schemas and enables multi-dataset pretraining for general-purpose GNNs, paving the way toward foundation models for relational graph data.