AI Paper Atlas

Research map / Graph learning / Knowledge and temporal graphs

Graph Convolutional Networks for Text Classification

Liang Yao, Chengsheng Mao, Yuan Luo

AAAI 2019 · Graph learning › Knowledge and temporal graphs · approach: Heterogeneous graph learning

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use a corpus-level graph with document-word and global word-word relations instead of the strongest listed baselineText GCN · 20NG testaccuracy: 0.8516 $\pm$ 0.0029 -> 0.8634 $\pm$ 0.0009Improved
Use a corpus-level graph with document-word and global word-word relations instead of the strongest listed baselineText GCN · R8 testaccuracy: 0.9699 $\pm$ 0.0012 -> 0.9707 $\pm$ 0.0010Improved
Use a corpus-level graph with document-word and global word-word relations instead of the strongest listed baselineText GCN · R52 testaccuracy: 0.9320 $\pm$ 0.0004 -> 0.9356 $\pm$ 0.0018Improved
Use a corpus-level graph with document-word and global word-word relations instead of the strongest listed baselineText GCN · Ohsumed testaccuracy: 0.6386 $\pm$ 0.0053 -> 0.6836 $\pm$ 0.0056Improved
Use a corpus-level graph instead of CNN with pretrained word embeddingsText GCN · MR testaccuracy: 0.7775 $\pm$ 0.0072 -> 0.7674 $\pm$ 0.0020Hurt
Use TF-IDF rather than term frequency alone for document-word edge weights (ablation)Text GCN · Text classificationaccuracyImproved
Use PMI rather than word co-occurrence counts for word-word edge weights (ablation)Text GCN · Text classificationaccuracyImproved
Use two GCN layers rather than one (ablation)Text GCN · Text classificationaccuracyImproved
Add more than two GCN layers (ablation)Text GCN · Text classificationaccuracyNo clear effect
Increase the sliding-window size from very small values toward 15 (ablation)Text GCN · R8 and MRaccuracyImproved
Increase the sliding-window size beyond 15 (ablation)Text GCN · R8 and MRaccuracyNo clear effect
Increase the first-layer embedding dimension from too-low values (ablation)Text GCN · R8 and MRaccuracyImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy