awslabs / awslabs/graphstorm

[Feature request] Support deterministic sampling for `GSgnnLinkPredictionTestDataLoader`

Open
#557 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
452
Forks
76
PR merge metrics
No merged PRs in 30d

Description

Validation/test sets for link-prediction tasks should yield the same set of negative links to make the metrics comparable across experiments on the same graph.

Contributor guide

Open the contributing guide

Research direction

Locate the GSgnnLinkPredictionTestDataLoader entry point and read how validation and test negative links are currently sampled. Determine how a deterministic sampling option should be exposed and verify that repeated runs on the same graph produce identical negative-link sets and comparable metrics.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
data, machine-learning
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.