随机选取40%-50%的表示值，能更高效地检测环境音频伪造。
Representation Loss Minimization with Randomized Selection Strategy for Efficient Environmental Fake Audio Detection
- 从基础模型表示中随机选40%-50%维度，替代传统降维方法
- 在参数减少一半以上时，检测性能不降反升
- 适合需要低延迟、轻量级音频伪造检测的应用场景
基础模型的适配显著推进了环境音频深度伪造检测（EADD）的发展。这些模型通常以微调或冻结状态用于下游任务。然而，其表示维度较高，导致下游模型参数量大，计算开销高。现有方法常采用SOTA无监督降维技术（如PCA、SVD、KPCA、GRP）压缩表示，以实现高效检测。但实验发现，此类方法会带来性能下降。本文揭示表示向量中存在冗余信息，仅随机选取40%-50%的表示值并构建下游模型，即可保持甚至提升性能。该策略在降低近一半模型参数和推理时间的同时，表现优于当前最优降维方法。
原文摘要 · Abstract (English)
The adaptation of foundation models has significantly advanced environmental audio deepfake detection (EADD), a rapidly growing area of research. These models are typically fine-tuned or utilized in their frozen states for downstream tasks. However, the dimensionality of their representations can substantially lead to a high parameter count of downstream models, leading to higher computational demands. So, a general way is to compress these representations by leveraging state-of-the-art (SOTA) unsupervised dimensionality reduction techniques (PCA, SVD, KPCA, GRP) for efficient EADD. However, with the application of such techniques, we observe a drop in performance. So in this paper, we show that representation vectors contain redundant information, and randomly selecting 40-50% of representation values and building downstream models on it preserves or sometimes even improves performance. We show that such random selection preserves more performance than the SOTA dimensionality reduction techniques while reducing model parameters and inference time by almost over half.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。