Research map / Probabilistic and scientific modeling / Dynamical systems
Learning Explicit Single-Cell Dynamics Using ODE Representations
Jan-Philipp von Bassewitz, Adeel Pervez, Marco Fumero, Matthew R Robinson, Theofanis Karaletsos, Francesco Locatello
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 |
|---|---|---|---|
| replace OT-based flow matching with Cell-MNN for interpolation (5D) | Cell-MNN · Single-cell interpolation EB/Multi | EMD: avg 0.776 (OT-MFM) -> 0.741 | Improved |
| compare to OT-MFM on Cite | Cell-MNN · Cite (5D) | EMD: 0.724 (OT-MFM) vs 0.791 | Hurt |
| force one eigenvalue of operator A to zero (ablation) | Cell-MNN · Single-cell interpolation (5D) | EMD average: 0.741 -> 0.748 | No clear effect |
| force one eigenvalue of operator A to zero (ablation) | Cell-MNN · TRRUST gene-interaction classification (EB) | F1: 60% -> 69% | Improved |
| amortized joint training on Cite and Multi vs OT-CFM | Cell-MNN · Cite+Multi amortized | EMD average: 0.835 (OT-CFM) -> 0.768 | Improved |
| use Cell-MNN instead of OT-CFM at 250k-cell scale | Cell-MNN · inflated EB/Cite/Multi (250,000 cells) | MMD | Improved |
| use Cell-MNN operator for GRN signs instead of OT-CFM Jacobian or SCODE | Cell-MNN (1EV=0) · TRRUST on EB | F1 average (%): 48 (OT-CFM J) -> 69 | Improved |
| scale to 10D PCA embedding versus OT-CFM | Cell-MNN · 10D PCA, average EMD | EMD: 1.592 (OT-CFM) vs 1.599 | No clear effect |
| scale to 50D/100D PCA | Cell-MNN · Multi 50D/100D | EMD: Multi 50D 43.926 vs OT-MFM 45.160; Cite 50D 38.803 vs OT-MFM 36.394 | Mixed |
Closest papers by meaning
- Joint Velocity-Growth Flow Matching for Single-Cell Dynamics Modeling (NeurIPS 2025 · similarity 0.85)
- Integrating GNN and Neural ODEs for Estimating Non-Reciprocal Two-Body Interactions in Mixed-Species Collective Motion (NeurIPS 2024 · similarity 0.85)
- Meta Flow Matching: Integrating Vector Fields on the Wasserstein Manifold (ICLR 2025 · similarity 0.85)
- GENOT: Entropic (Gromov) Wasserstein Flow Matching with Applications to Single-Cell Genomics (NeurIPS 2024 · similarity 0.84)
- CellPLM: Pre-training of Cell Language Model Beyond Single Cells (ICLR 2024 · similarity 0.84)
- Reconstructing Cell Lineage Trees from Phenotypic Features with Metric Learning (ICML 2025 · similarity 0.83)
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.
