Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Active Timepoint Selection for Learning Measure-Valued Trajectories
Nicolas Huynh, Mihaela van der Schaar
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 |
|---|---|---|---|
| use active uncertainty-based timepoint selection instead of uniform grid (branch duration 0.05) | LOT+GP surrogate · Synthetic branching | Rel. W2: 0.231 | Improved |
| widen branching windows to 0.20 when comparing active vs uniform (ablation) | Active selection · Synthetic branching | Rel. W2 vs Uniform: 0.231 -> -0.172 | Hurt |
| active selection vs uniform at branch duration 0.20 on velocity-weighted metric (ablation) | Active selection · Synthetic branching | Rel. w-W2 vs Uniform: 0.357 -> 0.071 | Improved |
| use Wasserstein barycenter reference instead of first snapshot (ablation) | LOT embedding · Synthetic | Mean abs. error: 0.2867 -> 0.0004 | Improved |
| use Wasserstein barycenter reference instead of first snapshot (ablation) | LOT embedding · Single-cell | Mean abs. error: 2.0529 -> 1.2057 | Improved |
| reduce PCA rank to K=2 (removing high-dimensional latent) (ablation) | Active LOT+GP · Synthetic and single-cell ablation | reconstruction error | Improved |
| add intrinsic time warping (ablation) | Active LOT+GP · Synthetic and single-cell ablation | reconstruction error | Improved |
| update reference to Wasserstein barycenter each iteration instead of fixing to first snapshot (ablation) | Active LOT+GP · Synthetic ablation | reconstruction error | Improved |
| use Matern-5/2 kernel instead of RBF (ablation) | Active LOT+GP · Synthetic and single-cell ablation | reconstruction error | No clear effect |
| use LOT embedding instead of moments or kernel mean embeddings (ablation) | Active selection · Synthetic and single-cell | reconstruction error | Improved |
| use Ours reconstruction instead of MMFM, 12 snapshots | Snapshots selected by active method · Synthetic | Mean W2: 1.746 -> 0.031 | Improved |
| use Ours reconstruction instead of MMFM, 20 snapshots | Snapshots selected by active method · Single-cell | Mean W2: 4.967 -> 2.906 | Improved |
Closest papers by meaning
- Active Target Discovery under Uninformative Priors: The Power of Permanent and Transient Memory (NeurIPS 2025 · similarity 0.85)
- Informed Initialization for Bayesian Optimization and Active Learning (NeurIPS 2025 · similarity 0.85)
- Robust Active Measuring under Model Uncertainty (AAAI 2024 · similarity 0.85)
- Transductive Active Learning: Theory and Applications (NeurIPS 2024 · similarity 0.84)
- ALINE: Joint Amortization for Bayesian Inference and Active Data Acquisition (NeurIPS 2025 · similarity 0.84)
- This Too Shall Pass: Removing Stale Observations in Dynamic Bayesian Optimization (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.
