Research map / Probabilistic and scientific modeling / Dynamical systems
Geometric Flow Grounding: A Unified Manifold Decoupling Framework for Dynamics Discovery and Verification
Chang Yu, Yuxuan Luo, Yixuan Du, Yuqing Zhou, Siyuan Li, Jingbo Zhou, jiawei jiang, Zhen Lei, Stan Z. Li
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 JVP-based tangent projection (ablation) | GFG · Pancreas | CBDir: 0.405 -> 0.588 | Improved |
| add JVP-based tangent projection (ablation) | GFG · Mouse Brain | CBDir: 0.465 -> 0.657 | Improved |
| add VQ velocity codebook (ablation) | GFG · Pancreas | CBDir: 0.125 -> 0.588 | Improved |
| add VQ velocity codebook (ablation) | GFG · Mouse Brain | CBDir: 0.106 -> 0.657 | Improved |
| add JVP tangent projection (ablation) | GFG · Mouse Brain | VeloCoh: 0.695 -> 0.864 | Improved |
| add JVP tangent projection (ablation) | GFG · Pancreas | VeloCoh: 0.993 -> 0.994 | No clear effect |
| add VQ codebook (ablation) | GFG · Pancreas | ICCoh: 0.972 -> 0.976 | No clear effect |
| set geometric weight lambda above 0 (ablation) | GFG · Pancreas | CBDir: 0.405 -> 0.588 | Improved |
| increase codebook size from 32 to 64 (ablation) | GFG · Pancreas | CBDir: 0.588 -> 0.580 | No clear effect |
| use GFG with tangent manifold constraints instead of SINDy | SINDy · 2D circle rotation ODE | one-step rollout MSE: 1.2902 -> 0.0003 | Improved |
| use GFG instead of IRK-SINDy | IRK-SINDy · 2D circle rotation ODE | VPT: 3.8 -> 26.1 | Improved |
| combine GFG with NSG-VD | NSG-VD · AI-generated video detection | Recall: 88.0 -> 93.2 | Improved |
Closest papers by meaning
- Markovian Dynamics Enforcer: Feasibility Preserving Correction on Learned Dyanmics Manifolds (NeurIPS 2026 · similarity 0.85)
- A Riemannian Framework for Learning Reduced-order Lagrangian Dynamics (ICLR 2025 · similarity 0.85)
- ManifoldCache: Training-Free Diffusion Acceleration via Constraint Manifold Caching (NeurIPS 2026 · similarity 0.84)
- Let's do the time-warp-attend: Learning topological invariants of dynamical systems (ICLR 2024 · similarity 0.84)
- Metric Flow Matching for Smooth Interpolations on the Data Manifold (NeurIPS 2024 · similarity 0.84)
- Rectifying Categorical Flows on Statistical Manifolds for One-Step Generation (NeurIPS 2026 · 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.
