超越语义准确率：面向安全关键型语言理解的后果感知评估

中文摘要 AI 辅助

语言模型在安全关键型操作中是否值得信赖？在这类场景下，语义指标上的出色表现并不能保证操作可靠性：误读高度、遗漏执行条件或混淆呼叫信号，可能在标准F1指标下得分良好，却会带来极具不对称性的操作后果。我们在空管（ATC）领域研究该问题，这里管制员与飞行员的通信要求近乎零误差，并采用后果感知评估来检验语义分数是否会误判操作可靠性。该框架被实例化为一个受控诊断性ATC基准，其基于航空标准及来自三个国家的40名空管人员的反馈。我们评估了8个模型，发现存在系统性的语义-安全差距：即使是在标准指标下看似可靠的模型，传统分数给出的性能估计也远高于后果感知评估。风险感知微调可缩小但无法消除该差距，这表明在任何实际安全关键型部署声明前，后果感知评估是标准NLP指标的必要补充。

英文摘要

Can language models be trusted in safety- critical operations? In such settings, strong per- formance on semantic metrics does not guaran- tee operational reliability: a misread altitude, a dropped execution condition, or a confused call- sign may score well under standard F1 yet carry sharply asymmetric operational consequences. We study this problem in air traffic control (ATC), where controller-pilot communication demands near-zero error tolerance, and use consequence-aware evaluation to test whether semantic scores misstate operational reliabil- ity. The framework is instantiated in a con- trolled diagnostic ATC benchmark grounded in aviation standards and feedback from 40 air traffic controllers across three countries. Evaluating 8 models, we uncover a system- atic semantic-safety gap: conventional scores give substantially higher performance estimates than consequence-aware evaluation, even for models that appear reliable under standard met- rics. Risk-aware fine-tuning narrows but does not close this gap, showing that consequence- aware evaluation is a necessary complement to standard NLP metrics before any real safety- critical deployment claim

发表机构

• ATMRI, Nanyang Technological University (NTU)（南洋理工大学ATMRI）
• Centre of AI Research, VinUniversity（文大学人工智能研究中心）
• School of Management, Indian Institute of Technology Mandi（印度理工大学曼迪分校管理学院）

机构由 AI 辅助整理，请以论文原文为准。