用一致性正则提升音频事件识别，小样本下效果更显著
Improving Audio Event Recognition with Consistency Regularization
- 对音频增强后版本强制模型预测一致，提升泛化能力
- 小数据集（2万条）上准确率提升明显，大样本（180万）也有效
- 半监督学习中用20万标注+180万无标注数据，性能超越纯监督
一致性正则化（CR）通过强制模型在数据增强后的不同视图间保持预测一致，已在自动语音识别中取得成效。本文首次将其应用于音频事件识别任务，并在AudioSet数据集上验证了有效性。针对小规模（约2万条）和大规模（约1.8百万条）的有监督训练集，我们进行了详尽消融实验，结果表明：即使在已大量使用数据增强的监督基线之上，CR仍能带来稳定提升；且更强的增强方式与多重增强策略在小样本场景下进一步增益显著。此外，我们将CR扩展至半监督设置，使用20万标注样本与180万未标注样本，所获模型性能优于仅用小规模标注数据训练的最佳模型。
原文摘要 · Abstract (English)
Consistency regularization (CR), which enforces agreement between model predictions on augmented views, has found recent benefits in automatic speech recognition [1]. In this paper, we propose the use of consistency regularization for audio event recognition, and demonstrate its effectiveness on AudioSet. With extensive ablation studies for both small ($\sim$20k) and large ($\sim$1.8M) supervised training sets, we show that CR brings consistent improvement over supervised baselines which already heavily utilize data augmentation, and CR using stronger augmentation and multiple augmentations leads to additional gain for the small training set. Furthermore, we extend the use of CR into the semi-supervised setup with 20K labeled samples and 1.8M unlabeled samples, and obtain performance improvement over our best model trained on the small set.
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。