Research map / Probabilistic and scientific modeling / Variational and generative sampling
Proppo: a Message Passing Framework for Customizable and Composable Learning Algorithms
Paavo Parmas, Takuma Seno
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 total propagation (step-wise inverse-variance weighting of LR and RP) instead of choosing the best of RP or LR | Proppo gradient estimators · Chaotic RNN, beta around 2.5 | gradient accuracy ratio: up to around 100 times | Improved |
| use reparameterization gradients instead of likelihood ratio at low inverse temperature | MC gradient estimators · Chaotic RNN, low beta | gradient variance | Improved |
| increase inverse temperature beta into chaotic regime for reparameterization gradients | RP gradient estimator · Chaotic RNN, beta around 2.5 | gradient variance | Hurt |
| increase horizon H from 10 to 100 for RP gradients | RP gradient estimator · Chaotic RNN | gradient variance | Hurt |
| replace step-wise TP with naive combination of separately computed LR and RP gradients (ablation) | Total propagation · Chaotic RNN | gradient accuracy ratio | Hurt |
| use TP in this RNN instead of prior cart-pole setting | Total propagation vs LR · Prior work comparison | improvement over LR: 3 times -> multiple orders of magnitude | Improved |
Closest papers by meaning
- Adaptive-Gradient Policy Optimization: Enhancing Policy Learning in Non-Smooth Differentiable Simulations (ICML 2024 · similarity 0.85)
- ATraDiff: Accelerating Online Reinforcement Learning with Imaginary Trajectories (ICML 2024 · similarity 0.84)
- Prior-Guided Diffusion Planning for Offline Reinforcement Learning (NeurIPS 2025 · similarity 0.84)
- GenPO: Generative Diffusion Models Meet On-Policy Reinforcement Learning (NeurIPS 2025 · similarity 0.84)
- Diffusion Spectral Representation for Reinforcement Learning (NeurIPS 2024 · similarity 0.84)
- TIC-GRPO: Provable and Efficient Optimization for Reinforcement Learning from Human Feedback (NeurIPS 2026 · 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.
