AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

Neural Bridge Sampling for Evaluating Safety-Critical Autonomous Systems

Aman Sinha, Matthew O'Kelly, Russ Tedrake, John C. Duchi

NeurIPS 2020 · Probabilistic and scientific modeling › Molecules and proteins · approach: Sparse concepts and hypothesis testing · 16 citations (OpenAlex)

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 changedWhereMeasuredEffect
use neural bridge sampling instead of naive Monte Carlorare-event estimation · Syntheticrelative mean-square error: 1.1821 -> 0.0051Improved
use neural bridge sampling instead of naive Monte Carlorare-event estimation · MountainCarrelative mean-square error: 0.2410 -> 0.0945Improved
use neural bridge sampling instead of naive Monte Carlorare-event estimation · Rocket2relative mean-square error: 0.0865 -> 0.0078Improved
use neural bridge sampling instead of naive Monte Carlorare-event estimation · WorldModelRacerrelative mean-square error: 0.9508 -> 0.1218Improved
add neural warping with normalizing flows to bridge sampling (ablation)bridge sampling · MountainCarrelative mean-square error: 0.3856 -> 0.0945Improved
add neural warping with normalizing flows to bridge sampling (ablation)bridge sampling · Syntheticrelative mean-square error: 0.0514 -> 0.0051Improved
add neural warping with normalizing flows to bridge sampling (ablation)bridge sampling · Rocket1relative mean-square error: 0.0129 -> 0.0102Improved
add neural warping with normalizing flows to bridge sampling (ablation)bridge sampling · AttentionAgentRacerrelative mean-square error: 0.9030 -> 0.2285Improved
use neural bridge sampling instead of adaptive multilevel splittingrare-event estimation · MountainCarrelative mean-square error: 0.5424 -> 0.0945Improved
use adaptive methods (AMS, B) instead of naive Monte Carlorare-event estimation · MountainCarrelative mean-square error: 0.2410 -> 0.5424 (AMS)Hurt
use adaptive multilevel splitting instead of naive Monte Carlorare-event estimation · Syntheticrelative mean-square error: 1.1821 -> 0.0162Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy