神经网络音频水印的向后兼容性研究：解决检测器升级导致误报问题
Backward Compatibility in Zero-Bit Neural Audio Watermarking
论文解决了一个很实际的问题：检测器升级后旧水印失效、误报飙升。用 legacy chain 加 ADL 损失来修复，方法挺清晰。
arXiv 论文 2610.22296 将零比特音频水印中的向后兼容性概念形式化，针对嵌入器与检测器频繁解耦更新的真实部署场景。作者提出一种训练框架，让新检测器与冻结的旧版嵌入器序列（legacy chain）配对训练，并使用性能加权采样保持对所有旧版本的检测精度。研究指出这种耦合会引发嵌入器坍缩效应，使对链外嵌入器的真负率（TNR）灾难性下降。论文随后提出 Angular Diversity Loss（ADL），强制每个新嵌入器占据输出空间的不同区域，最终在检测精度和感知音质无损失的情况下恢复接近满分的 TNR。
Backward Compatibility in Zero-Bit Neural Audio Watermarking
Neural watermarking systems are typically trained as tightly coupled embedder-detector pairs and implicitly assume static deployment. In real-world deployments, embedders and detectors are frequently updated and selectively decoupled for security and versioning purposes, which can break detection in previously embedded content and, more critically, increase false-positive rates on embedders that a given detector is intentionally not intended to recognize. In this paper the concept of backward compatibility in zero-bit audio watermarking is formalized. A training framework is proposed that pairs a new detector with a sequence of frozen legacy embedders (its legacy chain), using a performance-weighted sampling strategy so that the detector retains accuracy across all earlier versions. This coupling, while necessary for backward compatibility, is shown to induce an embedder-collapse effect that catastrophically degrades the true negative rate (TNR) against out-of-chain embedders. To counteract this, a Angular Diversity Loss (ADL) is introduced that explicitly forces each new embedder to occupy a distinct region of the output space. The combined approach restores near-perfect TNR with no measurable cost to detection accuracy or perceptual audio quality.
