Research map / Graph learning / Graph self-supervised learning
Joint Label Inference in Networks
Deepayan Chakrabarti, Stanislav Funiak, Jonathan Chang, Sofus A. Macskassy
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add group-membership nodes to EdgeExplain graph (ablation) | EdgeExplain (REL) · Facebook network, K=100 | recall at 3: maximum lift of only 1.2% for employer | No clear effect |
| add group-membership nodes to EdgeExplain graph (ablation) | EdgeExplain (REL) · Facebook network, K=100 | recall at 1: -2.8% employer, 0.4% current city | Mixed |
| increase number of retained friends K beyond a point (20 to 50, 100, 200, 400) (ablation) | EdgeExplain · Facebook network | recall lift over K=20 | Mixed |
| use REL inference instead of VAR (ablation) | EdgeExplain · Facebook network, K=100 | lift in recall over LP | Improved |
| use EdgeExplain instead of label propagation | Label inference · Facebook network | recall | Improved |
| use EdgeExplain instead of label propagation on sparse data | Label inference · Google+ network | recall at 3 | No clear effect |
| use REL instead of LP in simulations | EdgeExplain · Simulated ego networks (N=1000, 50 repetitions) | accuracy | Improved |
| use VAR instead of REL in simulations when most friends' labels are known with high confidence (ablation) | EdgeExplain · Simulated ego networks | accuracy | Mixed |
| clip label distributions to top 8 entries (ablation) | EdgeExplain · Facebook network | accuracy and running time: top 8 entries | No clear effect |
Closest papers by meaning
- Multi-Label Node Classification with Label Influence Propagation (ICLR 2025 · similarity 0.87)
- TEG-DB: A Comprehensive Dataset and Benchmark of Textual-Edge Graphs (NeurIPS 2024 · similarity 0.84)
- Pairwise Alignment Improves Graph Domain Adaptation (ICML 2024 · similarity 0.84)
- HGDL: Heterogeneous Graph Label Distribution Learning (NeurIPS 2024 · similarity 0.84)
- ML-GOOD: Towards Multi-Label Graph Out-Of-Distribution Detection (AAAI 2025 · similarity 0.84)
- GLDL: Graph Label Distribution Learning (AAAI 2024 · similarity 0.84)
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.
