Research map / Graph learning / Graph self-supervised learning
DisenCite: Graph-Based Disentangled Representation Learning for Context-Specific Citation Generation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use context-typed citation graph information and disentangled representations instead of the strongest graph baseline | Citation text generation on GCite · GCite test | BLEU-1: 0.5252 -> 0.5418 | Helped |
| Use context-typed citation graph information and disentangled representations instead of the strongest baseline | Citation position prediction on GCite · GCite test | Macro-F1: 0.7675 -> 0.7835 | Helped |
| Add intra-MI estimation between same-aspect paper representations and neighbors (ablation) | DisenCite · GCite | ROUGE-L | Helped |
| Add inter-MI estimation to separate different aspect representations (ablation) | DisenCite · GCite | ROUGE-L | Helped |
| Add intra-MI estimation between same-aspect paper representations and neighbors (ablation) | DisenCite · GCite | Macro-F1 | Hurt |
| Add inter-MI estimation to separate different aspect representations (ablation) | DisenCite · GCite | Macro-F1 | Hurt |
| Add both intra-MI and inter-MI estimation (ablation) | DisenCite · GCite | ROUGE-L | Helped |
| Propagate paper states through citation-graph transformer layers rather than using zero layers (ablation) | DisenCite · GCite | Macro-F1 | Helped |
| Increase graph-transformer depth beyond one layer (ablation) | DisenCite · GCite | Macro-F1 | No clear effect |
Closest papers by meaning
- HLM-Cite: Hybrid Language Model Workflow for Text-based Scientific Citation Prediction (NeurIPS 2024 · similarity 0.86)
- SelfCite: Self-Supervised Alignment for Context Attribution in Large Language Models (ICML 2025 · similarity 0.82)
- COMPOSE: Composing Future Theorems from Citations and Formal Structure (NeurIPS 2026 · similarity 0.82)
- Graph-based Unsupervised Disentangled Representation Learning via Multimodal Large Language Models (NeurIPS 2024 · similarity 0.81)
- Chain-of-Thought Improves Text Generation with Citations in Large Language Models (AAAI 2024 · similarity 0.80)
- DTGB: A Comprehensive Benchmark for Dynamic Text-Attributed Graphs (NeurIPS 2024 · similarity 0.80)
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.