---
title: "Detecting Silent Action-Interface Mismatch with Calibrated Relational World Models"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2611.06653v1"
permalink: "https://feed.astralor.com/items/itm_261a5d57bd20633a64844dff"
published: "2026-11-04T15:30:52.000Z"
score: 59
channel: "ai"
tags: ["安全","Detecting Silent Action-Interface","Mismatch","Calibrated Relational World","Models","Clean","LIBERO Spatial","Gaussian"]
---

# Detecting Silent Action-Interface Mismatch with Calibrated Relational World Models

## At a glance

A robot action can satisfy shape and range constraints yet carry the wrong semantics: axes may be permuted, signs inverted, rotations rescaled, or gripper polarity reversed. We study learned state-action-effect relation…

## Key points

- A robot action can satisfy shape and range constraints yet carry the wrong semantics: axes may be permuted, signs inverted, rotations rescaled, or gripper polarity reversed.
- We study learned state-action-effect relations as offline contracts for detecting these mismatches and ranking their likely fields.
- Clean held-out transitions calibrate alarms; a fixed inverse vocabulary supports closed-set localization.

## Full text

A robot action can satisfy shape and range constraints yet carry the wrong semantics: axes may be permuted, signs inverted, rotations rescaled, or gripper polarity reversed. We study learned state-action-effect relations as offline contracts for detecting these mismatches and ranking their likely fields. Clean held-out transitions calibrate alarms; a fixed inverse vocabulary supports closed-set localization. On ten LIBERO Spatial tasks comprising 432 source episodes, a capacity-controlled comparison at 2.5% empirical clean false-positive rate gives 46.1% recall for a ridge forward residual, 61.0% for a four-component joint Gaussian mixture, and 65.7% for the exact conditional of that same mixture. The joint score retains a 14.9 percentage-point advantage over ridge, while conditioning improves recall by a further 4.7 points. Joint scoring instead has higher localization: 95.8% versus 94.6%. These matched-rate comparisons are retrospective diagnostics, not deployment calibration guarantees. Earlier Object/Goal results broaden tasks within the LIBERO convention; DROID, all three HYDRA tasks, short temporal faults, small offsets, exact repair, and simulated mitigation fail their respective gates, while xArm contributes one task only. The evidence supports a bounded offline diagnostic and a detection-localization tradeoff, rather than joint-density superiority or cross-robot safety.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_261a5d57bd20633a64844dff