link-prediction

GNN-based prediction of connections in a spatial configuration (floor plan).

https://github.com/metis-caad/link-prediction

Science Score: 52.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • CITATION.cff file
    Found CITATION.cff file
  • codemeta.json file
    Found codemeta.json file
  • .zenodo.json file
    Found .zenodo.json file
  • DOI references
  • Academic publication links
  • Academic email domains
  • Institutional organization owner
    Organization metis-caad has institutional domain (www.ar.tum.de)
  • JOSS paper metadata
  • Scientific vocabulary similarity
    Low similarity (7.6%) to scientific vocabulary
Last synced: 6 months ago · JSON representation ·

Repository

GNN-based prediction of connections in a spatial configuration (floor plan).

Basic Info
  • Host: GitHub
  • Owner: metis-caad
  • License: apache-2.0
  • Language: Python
  • Default Branch: master
  • Homepage:
  • Size: 13.5 MB
Statistics
  • Stars: 3
  • Watchers: 1
  • Forks: 1
  • Open Issues: 0
  • Releases: 6
Created about 4 years ago · Last pushed over 2 years ago
Metadata Files
Readme License Citation

README.md

link-prediction

AI-based autocompletion of graph-based architectural spatial configurations using link prediction with graph neural networks (GNN).

The main goal of the approach is to estimate the probability of relations between the rooms of the spatial configuration graph using the relevant semantic information.

The approach is published in the context of the paper "Autocompletion of Design Data in Semantic Building Models using Link Prediction and Graph Neural Networks" submitted and presented @ eCAADe 2022.

Requirements

Following packages are required to run link prediction:

python3 & python3-pip

django & djangorestframework

Deep Graph Library (DGL) (Non-CUDA version should suffice)

Matplotlib

Run evaluation

Run ./train.sh in console.

You should get results similar to the evaluation in the paper (see linked paper).

The evaluation was tested on Ubuntu 22.04 LTS with Python 3.10 and CUDA 11.6.

Run API (lpapi)

Run ./lpapi.sh in console.

Owner

  • Name: metis
  • Login: metis-caad
  • Kind: organization

CAAD+AI research project

Citation (CITATION.cff)

cff-version: 1.2.0
message: "If you use this software, please cite it as below."
authors:
  - family-names: Eisenstadt
    given-names: Viktor
    orcid: https://orcid.org/0000-0001-6567-0943
  - family-names: Bielski
    given-names: Jessica
    orcid: https://orcid.org/0000-0003-4936-1993
  - family-names: Langenhan
    given-names: Christoph
    orcid: https://orcid.org/0000-0002-6922-2707
  - family-names: Karaali
    given-names: Ozan
    orcid: 
  - family-names: Althoff
    given-names: Klaus-Dieter
    orcid: https://orcid.org/0000-0002-7330-6540
  - family-names: Dengel
    given-names: Andreas
    orcid: https://orcid.org/0000-0002-6100-8255
title: "link-prediction"
version: 1.0.2
doi: 10.5281/zenodo.8139947
URL: https://github.com/metis-caad/link-prediction
date-released: 2023-07-12

GitHub Events

Total
Last Year