ArtifactNet：通过刑侦物理检测AI生成音乐
ArtifactNet: Detecting AI-Generated Music via Forensic Residual Physics
AI总结：
ArtifactNet通过刑侦物理方法检测AI生成音乐，利用编码器残留特征进行分类，相比其他方法在准确率和误报率上更优。
摘要：
我们提出了ArtifactNet，一个轻量级框架，通过将问题重新表述为刑侦物理——提取和分析神经音频编码器不可避免地印在生成音频上的物理残留物。一个有界遮罩的UNet（ArtifactUNet，3.6M参数）从幅度频谱图中提取编码器残留物，然后通过HPSS分解为7通道的刑侦特征，再通过紧凑的CNN（0.4M参数；总计4.0M参数）进行分类。我们引入了ArtifactBench，一个包含6,183首曲目的多生成器评估基准（4,383首AI生成音乐来自22个生成器，1,800首真实音乐来自6个不同来源）。每首曲目都标记有bench_origin以进行公平的零样本评估。在未见过的测试部分（n=2,263），ArtifactNet实现了F1=0.9829，FPR=1.49%，相比CLAM（F1=0.7576，FPR=69.26%）和SpecTTTra（F1=0.7713，FPR=19.43%）在相同条件下评估的公开检查点。编码器感知训练（4种WAV/MP3/AAC/Opus增强）进一步将跨编码器概率漂移降低了83%（Delta=0.95->0.16），解决了主要的编码器不变性失败模式。这些结果确立了刑侦物理——直接提取编码器级残留物——作为比表示学习更通用、参数更高效的方法来检测AI音乐，参数数量比CLAM少49倍，比SpecTTTra少4.8倍。
英文摘要：
Detecting artificially generated music requires distinguishing synthesis-related traces from musical content and artifacts introduced by audio distribution. We present ArtifactNet, a compact framework based on learned forensic residuals. A pretrained music source separator supplies residual targets during training, but the final system replaces it with ArtifactUNet, a task-directed bounded-mask extractor. Seven channels describe harmonic and percussive residual structure, their balance, and temporal variation, and a lightweight convolutional classifier produces recording scores. The extractor and classifier contain 4.03 million parameters. We evaluate the retained checkpoints on a frozen, recording-level test containing 534 generated and 479 real recordings after conservative source-family and identity exclusions from the reconstructed training lineage. At the pre-established threshold of 0.225, ArtifactNet obtains F1=0.9671, 93.63% recall, and no observed false positive; AUROC and average precision are 0.9982 and 0.9988. Released SpecTTTra and CLAM checkpoints are evaluated on the same files under their model-specific frontends. A diagnostic further shows that the released CLAM head changes substantially with execution batch composition. In a separate paired four-codec study on 100 real and 100 generated recordings, codec-aware extraction reduces the mean score range from 0.276 to 0.033 for real music but increases it from 0.035 to 0.249 for generated music; Opus recall falls from 96% to 74%. The results support compact residual detection while exposing class-dependent codec and implementation sensitivities.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
