Research map / Optimization / Optimizers
Trajectory Inference via Mean-field Langevin in Path Space
Lénaïc Chizat, Stephen Zhang, Matthieu Heitz, Geoffrey Schiebinger
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 grid-free MFL dynamics instead of gWOT on fixed support | Global Waddington-OT · bifurcating SDE with few samples per intermediate timepoint (N small) | RMS Energy Distance | Improved |
| use grid-free MFL dynamics instead of gWOT with many samples | Global Waddington-OT · bifurcating SDE, large N | RMS Energy Distance | No clear effect |
| reduce intermediate samples from N=64 to N=1 | MFL dynamics · bifurcating SDE | reconstruction quality | No clear effect |
| reduce intermediate samples from N=64 to N=1 | gWOT · bifurcating SDE | reconstruction quality | Hurt |
| account for branching (mass variation) in MFL dynamics (ablation) | MFL dynamics · bistable process | fraction of paths ending in lower branch: 77 vs ground truth 72 | Improved |
| neglect branching in MFL dynamics (ablation) | MFL dynamics · bistable process | fraction of paths ending in lower branch | Hurt |
| apply MFL dynamics to subsampled snapshots | subsampled snapshots · iPSC reprogramming, days 2.5-6 | Energy Distance to full dataset | Improved |
| MFL vs gWOT with best swept parameters | gWOT · iPSC reprogramming | Energy Distance to full dataset | Improved |
| scale step size and data-fitting bandwidth with temperature during simulated annealing (ablation) | MFL dynamics · simulated example | optimisation speed | Improved |
Closest papers by meaning
- Partially Observed Trajectory Inference using Optimal Transport and a Dynamics Prior (ICLR 2025 · similarity 0.92)
- Trajectory Inference with Smooth Schrödinger Bridges (ICML 2025 · similarity 0.91)
- Langevin-Informed Transfer Learning: Replacing the Target Samples by Black-Box Feedback (NeurIPS 2026 · similarity 0.87)
- Doob's Lagrangian: A Sample-Efficient Variational Approach to Transition Path Sampling (NeurIPS 2024 · similarity 0.86)
- Statistical Spatially Inhomogeneous Diffusion Inference (AAAI 2024 · similarity 0.86)
- Mean-field Underdamped Langevin Dynamics and its Spacetime Discretization (ICML 2024 · similarity 0.86)
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.
