Research map / Graph learning / Knowledge and temporal graphs
Adaptive Memory Retention in Dynamic Graphs
Fabrizio De Castelli, Alessio Gravina, Moshe Eliasof, Carola-Bibiane Schönlieb, Davide Bacciu
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 weight constraints and antisymmetric parameterization (non-dissipative core) (ablation) | LAMP · ColoredLeafCounting three-color T=200 | test MSE: 1.327 -> 0.057 | Improved |
| add weight constraints and antisymmetric parameterization (non-dissipative core) (ablation) | LAMP · ColoredLeafCounting binary T=600 | test MSE: 1.373 -> 0.831 | Improved |
| add antisymmetric aggregation (beta not 0) (ablation) | LAMP without weight constraints · ColoredLeafCounting three-color T=600 | test MSE: 0.718 -> 0.650 | Mixed |
| learn dissipation Gamma instead of fixing it (ablation) | LAMP · ColoredLeafCounting three-color T=400 | test MSE: 0.523 -> 0.264 | Improved |
| learn dissipation Gamma instead of fixing it (ablation) | LAMP · ColoredLeafCounting binary T=600 | test MSE: 1.135 -> 0.831 | Improved |
| add learned dissipation Gamma over Gamma=0 (ablation) | LAMP · ColoredLeafCounting binary T=600 | test MSE: 0.983 -> 0.831 | Improved |
| add learned dissipation Gamma over Gamma=0 (ablation) | LAMP · ColoredLeafCounting three-color T=200 | test MSE: 0.051 -> 0.057 | No clear effect |
| add learned dissipation Gamma over Gamma=0 (ablation) | LAMP · ColoredLeafCounting binary T=200 | test MSE: 0.438 -> 0.467 | No clear effect |
| use LAMP instead of DCRNN | LAMP vs DCRNN · ColoredLeafCounting three-color T=200 | test MSE: 1.283 -> 0.057 | Improved |
| use LAMP instead of T-GCN | LAMP vs T-GCN · ColoredLeafCounting binary T=400 | test MSE: 0.764 -> 0.394 | Improved |
| use LAMP instead of GC-LSTM | LAMP vs GC-LSTM · AS-733 link prediction | F1: 91.22 -> 96.20 | Improved |
| use LAMP instead of DynGESN | LAMP vs DynGESN · ChickenPox Hungary | test MSE: 0.907 -> 0.748 | Improved |
Closest papers by meaning
- A Dynamical Systems-Inspired Pruning Strategy for Addressing Oversmoothing in Graph Attention Networks (ICML 2025 · similarity 0.87)
- Delay-DSGN: A Dynamic Spiking Graph Neural Network with Delay Mechanisms for Evolving Graph (ICML 2025 · similarity 0.87)
- A Deep Probabilistic Framework for Continuous Time Dynamic Graph Generation (AAAI 2025 · similarity 0.87)
- DyG-Mamba: Continuous State Space Modeling on Dynamic Graphs (NeurIPS 2025 · similarity 0.87)
- PRES: Toward Scalable Memory-Based Dynamic Graph Neural Networks (ICLR 2024 · similarity 0.86)
- TMetaNet: Topological Meta-Learning Framework for Dynamic Link Prediction (ICML 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.
