NeST-Bayes: Trust Inference for Multi-Robot Systems Without Prior Trust
DukeUniversity · hf · 2026-10-15
Duke researchers introduce NeST-Bayes, a neural-symbolic framework that jointly infers robot and hypothesis trust in multi-robot systems using only inter-robot observation consistency — no preassigned trust, reputation data, or side-channel identity signals required. Evidence scores from ego-centric relational graphs accumulate via Bayesian updates of Beta-distributed trust beliefs. It achieves the best object precision and robot recall across benchmarks, sustaining 98.8% robot precision / 93.8% recall under aggressive optimized attacks, and 92.5% robot precision in an occlusion-rich Webots warehouse.
More from Embodied
- LEGO builds an omnidirectional vehicle for automated warehouses — glenbeer · 2026-10-15
- SlipSense: multimodal tactile slip detection hits 96.7% F1 at 23.1ms with zero-shot robot transfer — chris_j_paxton · 2026-10-15
- BlackBerry launches QNX Launchpad to court early-stage Physical AI startups — pdamodaran · 2026-10-15
- Microduck robot runs onboard SLAM with just 64 pixels of depth, 4.8 cm median error — chris_j_paxton · 2026-10-15
- Dynamic Creatures exits stealth as Boston Dynamics' official entertainment and hospitality partner — PeterBowdenLive · 2026-10-15
- Figure AI valued at $39B tops humanoid chart, exceeds any listed robot maker — LexSokolin · 2026-10-15