Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Probabilities Are Not Enough: Formal Controller Synthesis for Stochastic Dynamical Models with Epistemic Uncertainty
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Account for uncertain mass instead of using a nominal-only abstraction | Drone reach-avoid controller · Drone reach-avoid, uncertain mass at robustness limit | controller safety: 100% safe controllers; baseline yields 6% unsafe controllers | Helped |
| Account for epistemic uncertainty | Building temperature control abstraction · Building temperature control, largest partition | runtime: 5 -> 8 | Hurt |
| Account for epistemic uncertainty | Building temperature control abstraction · Building temperature control, largest partition | transitions: 52 -> 82 million transitions | Mixed |
| Refine the state-space partition | Proposed controller synthesis method · Building temperature control | robust lower bound on reach-avoid probability | Helped |
| Decouple room dynamics to capture room interactions as bounded uncertainty | Building temperature control method · Building temperature control | scalability | Helped |
Closest papers by meaning
- Sampling-Based Robust Control of Autonomous Systems with Non-Gaussian Noise (AAAI 2022 · similarity 0.89)
- Safely Learning Controlled Stochastic Dynamics (NeurIPS 2025 · similarity 0.86)
- Robust Active Measuring under Model Uncertainty (AAAI 2024 · similarity 0.84)
- UnHiPPO: Uncertainty-aware Initialization for State Space Models (ICML 2025 · similarity 0.84)
- Near-Optimal Policy Identification in Robust Constrained Markov Decision Processes via Epigraph Form (ICLR 2025 · similarity 0.84)
- Sampling-Based Safe Reinforcement Learning (NeurIPS 2026 · similarity 0.83)
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.