microsoft/Graphormer

About Periodic boundary condition invariance

Open

#137 opened on Sep 22, 2022

View on GitHub
 (0 comments) (0 reactions) (0 assignees)Python (376 forks)auto 404
good first issue

Repository metrics

Stars
 (2,459 stars)
PR merge metrics
 (PR metrics pending)

Description

For general MPNN as CGCNN and Schnet, crystal graph as multi-graph is invariant under different supercell. I wonder if Graphormer can achieve the same invariance. More specifically, will the results of a unit cell input structure and a supercell structure of Graphormer keep the same? I don't think it achieve pbc invariance as the graph is changed for different supercell. I'm surprised that Graphormer can performs excellent not only molecules but also crystals! Thanks.

Contributor guide

Research direction
Investigate the periodic boundary condition invariance of Graphormer for crystal structures by reviewing the paper and code, and compare with MPNNs like CGCNN and Schnet to understand if Graphormer's attention mechanism respects PBC invariance.
Tech stack
python
Domain
ai
Issue type
Research
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.
Under 1 hour
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Fresh
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Mostly clear
Prerequisites
machine learninggraph neural networkscrystal structures
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
50