AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Text Classification with Heterogeneous Information Network Kernels

Chenguang Wang, Yangqiu Song, Haoran Li, Ming Zhang, Jiawei Han

AAAI 2016 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 82 citations (OpenAlex)

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
replace BOW features with HIN-links features (entities+relations)Naive Bayes · GCAT-SIMaccuracy: 88.49% -> 90.02%Improved
replace BOW features with HIN-links featuresSVM · GCAT-DIFaccuracy: 88.98% -> 91.19%Improved
add entities as flat features to BOWSVM · 20NG-DIFaccuracy: 96.66% -> 96.90%No clear effect
use all meta-paths (DWD+MP) instead of only DWD meta-path in KnowSim kernel (ablation)IndefSVMHIN+KnowSim · GCAT-SIMaccuracy: 95.63% -> 98.10%Improved
use indefinite-kernel SVM instead of zeroing negative eigenvalues (ablation)SVMHIN+KnowSim DWD+MP · GCAT-DIFaccuracy: 91.88% -> 93.51%Improved
KnowSim kernel with DWD meta-path (normalized) vs BOW SVM (ablation)SVM · 20NG-SIMaccuracy: 90.81% -> 92.32%Improved
IndefSVM KnowSim DWD+MP vs word embedding average SVMSVM (WEAvg) · GCAT-SIMaccuracy: 96.81% -> 98.10%Improved
SVMHIN vs SVM WEAvgSVMHIN · 20NG-DIFaccuracy: 98.27% (WEAvg) vs 97.20% (SVMHIN)No clear effect
use knowledge-graph GSim similarity instead of KnowSim in indefinite SVM (ablation)IndefSVM · 20NG-SIMaccuracy: 93.38% -> 50.44%Hurt

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