Research map / Reinforcement learning / Policy gradient and actor-critic
Reinforced Sequential Monte Carlo for Amortised Sampling
Sanghyeok Choi, Sarthak Mittal, Víctor Elvira, Jinkyoo Park, Esmeralda S. Whitammer
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 importance-weighted buffer (IW-Buf) (ablation) | TB on-policy · GMM40 (d=2) | Sinkhorn: 607.31 -> 6.50 | Improved |
| add SMC as behaviour policy (ablation) | TB on-policy · GMM40 (d=2) | Sinkhorn: 607.31 -> 39.99 | Improved |
| add importance-weighted buffer (IW-Buf) (ablation) | TB on-policy · ManyWell (d=32) | EUBO: 252.37 -> 166.30 | Improved |
| add importance-weighted buffer (IW-Buf) (ablation) | TB/SubTB + SMC · Robot4 (d=10) | Sinkhorn: 64.48 -> 0.39 | Improved |
| add importance-weighted buffer (IW-Buf) (ablation) | TB/SubTB + SMC · GMM40 (d=5) | ELBO: -10.56 -> -11.46 | Hurt |
| add importance-weighted buffer (IW-Buf) (ablation) | TB on-policy · GMM40 (d=50) | Sinkhorn: 3903.95 -> 4284.49 | Hurt |
| add importance-weighted buffer (IW-Buf) (ablation) | TB on-policy · ManyWell (d=64) | MMD: 0.243 -> 0.058 | Improved |
| add SMC as behaviour policy without buffer (ablation) | TB on-policy · Robot4 (d=10) | Sinkhorn: 1.72 -> 64.48 | Hurt |
| use IW-Buf instead of uniform buffer (ablation) | TB on-policy + buffer · ManyWell (d=32) | EUBO: 170.06 -> 166.30 | Improved |
| use IW-Buf instead of loss-prioritised buffer (ablation) | TB on-policy + buffer · ManyWell (d=32) | EUBO: 173.26 -> 166.30 | Improved |
| add learnt correction to learnt interpolation of intermediate targets (ablation) | SMC + IW-Buf · GMM40 (d=50) | Sinkhorn: 5091.1 -> 3579.2 | Improved |
| add learnt correction to linear interpolation (ablation) | SMC + IW-Buf · GMM40 (d=50) | Sinkhorn: x -> 4675.9 | Improved |
Closest papers by meaning
- Adaptive teachers for amortized samplers (ICLR 2025 · similarity 0.89)
- Reverse Diffusion Sequential Monte Carlo Samplers (NeurIPS 2025 · similarity 0.88)
- Improved off-policy training of diffusion samplers (NeurIPS 2024 · similarity 0.87)
- On scalable and efficient training of diffusion samplers (NeurIPS 2025 · similarity 0.87)
- Sequential Controlled Langevin Diffusions (ICLR 2025 · similarity 0.87)
- Scalable Discrete Diffusion Samplers: Combinatorial Optimization and Statistical Physics (ICLR 2025 · 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.
