Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Robust Generalized Schr\"{o}dinger Bridge via Sparse Variational Gaussian Processes
Minyoung Kim
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 deterministic spline path optimization with sparse variational GP posterior | GSBM · LiDAR (noisy stage cost) | CondSOC objective: 8506.1 -> 8300.0 | Improved |
| replace spline path optimization with GP posterior | GSBM · LiDAR (noise-free) | CondSOC objective: 6199.3 -> 5925.0 | Improved |
| replace spline path optimization with GP posterior | GSBM · Stunnel (uncertain obstacles) | CondSOC objective: 502.20 -> 452.30 | Improved |
| replace spline path optimization with GP posterior | GSBM · GMM (uncertain obstacles) | CondSOC objective: 101.6 -> 89.2 | Improved |
| replace spline path optimization with GP posterior | GSBM · Stunnel (deterministic obstacles) | CondSOC objective: 492.94 -> 488.78 | No clear effect |
| replace spline path optimization with GP posterior | GSBM · AFHQ dog to cat generation | FID: 12.39 -> 10.21 | Improved |
| use polynomial kernel instead of squared exponential (ablation) | GP-GSBM · LiDAR and Stunnel (deterministic) | CondSOC objective: LiDAR noise-free 5925.0 -> 6604.8; Stunnel 488.78 -> 556.50 | Hurt |
| use polynomial kernel instead of squared exponential (ablation) | GP-GSBM · Stunnel (uncertain obstacles) | CondSOC objective: 452.30 -> 435.37 | Improved |
| use stream-level GP instead of CondSOC likelihood GP | GP path model · LiDAR (noisy) | CondSOC objective: 8300.0 -> 12679.1 | Hurt |
| vary number of inducing points (ablation) | GP-GSBM · Stunnel/LiDAR | CondSOC objective | No clear effect |
Closest papers by meaning
- Generalized Schrödinger Bridge Matching (ICLR 2024 · similarity 0.90)
- Go With the Flow: Fast Diffusion for Gaussian Mixture Models (NeurIPS 2025 · similarity 0.87)
- Momentum Multi-Marginal Schrödinger Bridge Matching (NeurIPS 2025 · similarity 0.87)
- Trajectory Inference with Smooth Schrödinger Bridges (ICML 2025 · similarity 0.86)
- Sparse Gaussian Processes: Structured Approximations and Power-EP Revisited (NeurIPS 2025 · similarity 0.85)
- Scalable Bayesian Optimization via Focalized Sparse Gaussian Processes (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.
