pyg-team/pytorch_geometric

Explanability for Heterogeneous Graphs

Open

#5,630 opened on Oct 7, 2022

View on GitHub
 (0 comments) (0 reactions) (0 assignees)Python (3,514 forks)batch import
0 - Priority P0explainfeaturehelp wanted

Repository metrics

Stars
 (19,985 stars)
PR merge metrics
 (Avg merge 35d 1h) (14 merged PRs in 30d)

Description

🚀 Heterogeneous Graph Explanation

Explainability support for heterogeneous graphs. There exist simple ways to adapt current explainability methods into heterogeneous graphs. For example, we can use gradient-based methods by taking gradient with respect to adjacencies (edge_index) of different message types. We can also adapt GNNExplainer by creating a mask corresponding to each of the edge_index for different message types.

Thanks for the good points@Padarn and @wsad1! I'll discuss offline and update this issue with more details for things like HGTConv in a few days.

Tasks:

  • Extension of the current explainability framework to work for HeteroData objects.
  • Incorporate captum-based explainability implementation (such as SHAP, DeepLIFT) for HeteroData.
  • Potential extension to hypergraph use cases (where the graph data is represented as nodes and a set of hyperedges which can connect more than 2 nodes).

Alternatives

No response

Additional context

No response

Contributor guide

Research direction
Investigate how to extend existing explainability methods (e.g., gradient based, GNNExplainer) to heterogeneous graphs by creating separate masks for each message type or edge type.
Tech stack
pythonpytorch
Domain
machine learningbackend
Issue type
Feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
2
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-2 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Needs investigation
Prerequisites
PythonPyTorch
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
40