Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Multi-Relational Graph Neural Network for Out-of-Domain Link Prediction

Asma Sattar, Georgios Deligiorgis, Marco Trincavelli, Davide Bacciu

TL;DR

This work establishes a novel class of challenging tasks for dynamic multi-relational graphs involving out-of-domain link prediction, where the relationship being predicted is not available in the input graph, and introduces a novel Graph Neural Network model, named GOOD, designed specifically to tackle the out-of-domain generalization problem.

Abstract

Dynamic multi-relational graphs are an expressive relational representation for data enclosing entities and relations of different types, and where relationships are allowed to vary in time. Addressing predictive tasks over such data requires the ability to find structure embeddings that capture the diversity of the relationships involved, as well as their dynamic evolution. In this work, we establish a novel class of challenging tasks for dynamic multi-relational graphs involving out-of-domain link prediction, where the relationship being predicted is not available in the input graph. We then introduce a novel Graph Neural Network model, named GOOD, designed specifically to tackle the out-of-domain generalization problem. GOOD introduces a novel design concept for multi-relation embedding aggregation, based on the idea that good representations are such when it is possible to disentangle the mixing proportions of the different relational embeddings that have produced it. We also propose five benchmarks based on two retail domains, where we show that GOOD can effectively generalize predictions out of known relationship types and achieve state-of-the-art results. Most importantly, we provide insights into problems where out-of-domain prediction might be preferred to an in-domain formulation, that is, where the relationship to be predicted has very few positive examples.

Multi-Relational Graph Neural Network for Out-of-Domain Link Prediction

TL;DR

This work establishes a novel class of challenging tasks for dynamic multi-relational graphs involving out-of-domain link prediction, where the relationship being predicted is not available in the input graph, and introduces a novel Graph Neural Network model, named GOOD, designed specifically to tackle the out-of-domain generalization problem.

Abstract

Dynamic multi-relational graphs are an expressive relational representation for data enclosing entities and relations of different types, and where relationships are allowed to vary in time. Addressing predictive tasks over such data requires the ability to find structure embeddings that capture the diversity of the relationships involved, as well as their dynamic evolution. In this work, we establish a novel class of challenging tasks for dynamic multi-relational graphs involving out-of-domain link prediction, where the relationship being predicted is not available in the input graph. We then introduce a novel Graph Neural Network model, named GOOD, designed specifically to tackle the out-of-domain generalization problem. GOOD introduces a novel design concept for multi-relation embedding aggregation, based on the idea that good representations are such when it is possible to disentangle the mixing proportions of the different relational embeddings that have produced it. We also propose five benchmarks based on two retail domains, where we show that GOOD can effectively generalize predictions out of known relationship types and achieve state-of-the-art results. Most importantly, we provide insights into problems where out-of-domain prediction might be preferred to an in-domain formulation, that is, where the relationship to be predicted has very few positive examples.
Paper Structure (16 sections, 9 equations, 3 figures, 3 tables)

This paper contains 16 sections, 9 equations, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Problem Definition
  4. Modeling Multi-Relational embeddings for Multi-Relational Graph
  5. Model Architecture
  6. Loss Function
  7. Link Prediction Objective
  8. Coefficient Disentanglement Objective
  9. Benchmarks and Experimental Analysis
  10. Datasets
  11. Evaluation Methodology
  12. Related Approaches
  13. Performance Comparison
  14. Conclusion and Future Work
  15. Implementation Details
  16. ...and 1 more sections

Figures (3)

  • Figure 1: A multi-relational graph with nodes of different types connected with known relations (left), that are used to predict an unknown relationship (right) in a link prediction task.
  • Figure 2: High-level architecture diagram. From top to bottom: each row represents node embeddings generated with respect to different contexts. From left to right: each row represents the effect of a sub-context inside a main context. Dashed lines represent the residual connections. MixAGG denotes the aggregation block.
  • Figure 3: Multi-relational out-of-domain Vs. Single-relation in-domain ROC-AUC results.