How realistic must simulated failures be? Low fidelity suffices with little data
mangahomanga · x · 2026-11-06
From the JHU robot RL paper thread (5/n): how realistic does a simulated failure need to look?
Analyzing visual fidelity at different data scales: with little data, low-fidelity simulation can match high-fidelity simulation; as data grows, the higher-fidelity version holds up better.
A practical cost hint for teams using real-to-sim failure data to train reward models: maximum fidelity isn't necessary on a tight budget.
Related event: JHU Study Compares Four Failure Data Sources for Robot Reward Models(7 posts)→
More from Embodied
- Wayve demos Learned Safety Envelope: safety as a learned task head, not a bolt-on checker — alexgkendall · 2026-11-06
- Modded MacBook Pro gets a working touchscreen — used to play Honor of Kings — xiaohu · 2026-11-06
- Roboticist shares Stanford-collab research directions and BARS talk — Fowe · 2026-11-06
- RAX From Harvard/NVIDIA/Berkeley Lifts Robot Success From 26% to 94% With Frozen Weights — jiqizhixin · 2026-11-06
- Open-Weights Robot-Use Harness: Qwen3.8 Scores 31.39, Beating GPT-6 Astra's 28.97 — aigclink · 2026-11-06
- Dev builds Salvo battleship game for dual-screen iPhone Duo with head-to-head and AI play — HankYeomans · 2026-11-06