Research map / Architectures and efficiency / RNNs and normalization
Memory-Augmented Dynamic Neural Relational Inference
Dong Gong, Frederic Z. Zhang, Javen Qinfeng Shi, Anton van den Hengel
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 relation and entity memory pools | DNRI · Physics simulation (3 particles, dynamic relations) | prediction MSE step #25 (x1e-4): 5.36 -> 4.56 | Improved |
| add relation and entity memory pools | DNRI · Physics simulation (3 particles, dynamic relations) | prediction MSE step #1 (x1e-4): 0.177 -> 0.126 | Improved |
| add relation and entity memory pools | DNRI · Physics simulation (3 particles, dynamic relations) | relation F1: 0.57 -> 0.68 | Improved |
| add relation and entity memory pools | DNRI · Physics simulation (3 particles, dynamic relations) | relation precision: 0.76 -> 0.91 | Improved |
| add relation and entity memory pools | DNRI · Physics simulation (3 particles, dynamic relations) | relation accuracy: 0.97 -> 0.98 | Improved |
| add dynamic relations (DNRI over NRI) | NRI · Physics simulation (3 particles, dynamic relations) | prediction MSE step #25 (x1e-4): 39.3 -> 5.36 | Improved |
| add RelMem in prior (plus EntMem in decoder) (ablation) | MemDNRI with only EntMem · CMU mocap #118 and #35 | trajectory MSE | Improved |
| add EntMem in decoder (plus RelMem in prior) (ablation) | MemDNRI with only RelMem · CMU mocap #118 and #35 | trajectory MSE | Improved |
| add memory augmentation | DNRI · CMU mocap #35 (stable walking) | trajectory MSE | Mixed |
| add memory augmentation | DNRI · CMU mocap #118 | trajectory MSE | Improved |
| add memory augmentation | DNRI · Basketball player trajectories | trajectory MSE | Improved |
Closest papers by meaning
- A Graph Dynamics Prior for Relational Inference (AAAI 2024 · similarity 0.86)
- Online Relational Inference for Evolving Multi-agent Interacting Systems (NeurIPS 2024 · similarity 0.86)
- A Framework for Inference Inspired by Human Memory Mechanisms (ICLR 2024 · similarity 0.85)
- TStruct: Learning Shared Temporal Structures for Long-Term Time-Series Forecasting (NeurIPS 2026 · similarity 0.85)
- A mathematical theory of balancing relational generalization and memorization (NeurIPS 2026 · similarity 0.85)
- Learning dynamics in linear recurrent neural networks (ICML 2025 · 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.
