Research map / Vision-language and multimodal / Cross-modal learning
Unlocking Multi-Modal Potentials for Link Prediction on Dynamic Text-Attributed Graphs
Yuanyuan Xu, Wenjie Zhang, Ying Zhang, Xuemin Lin, Xiwei Xu
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 temporal modality modeling (ablation) | MoMent · Inductive link prediction (Enron, Stack elec, Stack ubuntu) | AUC: average improvement of 5.91% | Improved |
| add textual modality modeling (ablation) | MoMent · Inductive link prediction (Enron, Stack elec, Stack ubuntu) | AUC: average improvement of 15.89% | Improved |
| add temporal and textual modalities to structural-only model (ablation) | MoMent · Inductive link prediction | AUC: up to 19.39% | Improved |
| add distribution-level alignment (ablation) | MoMent · Inductive link prediction | AUC: up to 4.12% | Improved |
| add instance-level alignment (ablation) | MoMent · Inductive link prediction | AUC: up to 3.54% | Improved |
| add the dual-domain alignment loss (ablation) | MoMent · Inductive link prediction | AUC: up to 6.03% | Improved |
| add textual modality modeling (ablation) | MoMent · Enron (inductive) | AUC: 0.8387 -> 0.9711 | Improved |
| add temporal modality modeling (ablation) | MoMent · Stack ubuntu (inductive) | AUC: 0.8534 -> 0.9599 | Improved |
| add distribution-level alignment (ablation) | MoMent · Stack elec (inductive) | AUC: 0.9294 -> 0.9482 | Improved |
| add instance-level alignment (ablation) | MoMent · Enron (inductive) | AUC: 0.9482 -> 0.9711 | Improved |
| use MoMent multi-modal model instead of TGNN baselines | MoMent · 7 DyTAG datasets, link prediction | AUC/AP: 7.80% inductive, 3.39% transductive | Improved |
| use MoMent instead of LLM-driven LKD4DyTAG | MoMent · Link prediction | AUC/AP: up to 7.03% AUC and 10.70% AP | Improved |
Closest papers by meaning
- DTGB: A Comprehensive Benchmark for Dynamic Text-Attributed Graphs (NeurIPS 2024 · similarity 0.93)
- Unifying Text Semantics and Graph Structures for Temporal Text-attributed Graphs with Large Language Models (NeurIPS 2025 · similarity 0.91)
- TEG-DB: A Comprehensive Dataset and Benchmark of Textual-Edge Graphs (NeurIPS 2024 · similarity 0.88)
- SSTAG: Structure-Aware Self-Supervised Learning Method for Text-Attributed Graphs (NeurIPS 2025 · similarity 0.87)
- Generalization Principles for Inference over Text-Attributed Graphs with Large Language Models (ICML 2025 · similarity 0.87)
- A Deep Probabilistic Framework for Continuous Time Dynamic Graph Generation (AAAI 2025 · similarity 0.86)
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.
