【论文速递】Beyond Answer Changes: Matched Same-Class Controls for Local Audio Evidence

英文标题：Beyond Answer Changes: Matched Same-Class Controls for Local Audio Evidence

arXiv编号：2610.37605
作者：Yunuo Wang, Maoyuan Chen, Yuanjin Huang
来源：arXiv
原文链接：https://arxiv.org/abs/2610.37605

中文摘要：
For audio-language models, an answer change after a local sound-event edit can reflect a response to the event's class, sensitivity to waveform rewriting, or both. We compare matched same-class and different-class replacements of a queried event. Both use the same interval, level estimate, donor fitting, and crossfade, while the same-class replacement preserves the correct answer. Across 600 synthetic soundscapes and three models from two architecture families, different-class replacements produce 49.7 to 60.9 percentage points more answer changes than same-class replacements. Same-class accuracy remains within 5.2 percentage points of original accuracy. Among trials with correct original and same-class answers, soundscape-weighted selective accuracy for the inserted class is 73.7 to 81.9%, averaging eligible trials within each soundscape and then weighting eligible soundscapes equally. This protocol tests how model answers respond to local changes in event class.

本文由华人空间 AI 论文助手自动抓取翻译，仅供学术交流。