Research map / Probabilistic and scientific modeling / Dynamical systems
Graph Switching Dynamical Systems
Yongtuo Liu, Sara Magliacane, Miltiadis Kofinas, Efstratios Gavves
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 dynamic graph-based interaction modeling to switching dynamics | REDSDS on ODE-driven Particle · ODE-driven Particle | Accuracy: 0.708$\pm$0.037 -> 0.794$\pm$0.030 | Improved |
| add dynamic graph-based interaction modeling to switching dynamics | REDSDS on Salsa Couple Dancing · Salsa Couple Dancing | Accuracy: 0.504$\pm$0.052 -> 0.569$\pm$0.065 | Improved |
| model interactions between multiple objects, without learning a dynamic graph | REDSDS on ODE-driven Particle · ODE-driven Particle | Accuracy: 0.708$\pm$0.037 -> 0.766$\pm$0.045 | Improved |
| increase particles from 3 to 10 while interactions increase (ablation) | GRASS · ODE-driven Particle | NMI: 0.528 -> 0.519 | Hurt |
| increase particles from 3 to 10 while keeping interactions fixed (ablation) | GRASS · ODE-driven Particle | NMI: 0.528 -> 0.524 | No clear effect |
| use a dataset with no object interactions (ablation) | GRASS · ODE-driven Particle without interactions | F1: 0.790 -> 0.826 | Improved |
| use a dataset with no object interactions (ablation) | MOSDS · ODE-driven Particle without interactions | F1: 0.757 -> 0.817 | Hurt |
| increase maximum mode count from 3 to 10 (ablation) | GRASS · ODE-driven Particle | NMI: 0.528 -> 0.527 | No clear effect |
Closest papers by meaning
- Benchmarking Structural Inference Methods for Interacting Dynamical Systems with Synthetic Data (NeurIPS 2024 · similarity 0.86)
- Learning system dynamics without forgetting (ICLR 2025 · similarity 0.86)
- Rethink GraphODE Generalization within Coupled Dynamical System (ICML 2025 · similarity 0.85)
- CausalDynamics: A large‐scale benchmark for structural discovery of dynamical causal models (NeurIPS 2025 · similarity 0.84)
- Learning the Topology and Behavior of Discrete Dynamical Systems (AAAI 2024 · similarity 0.84)
- Neural Persistence Dynamics (NeurIPS 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.
