提出无样本增量学习方法AFT，解决环境声音识别中旧知识遗忘问题。
AFT: An Exemplar-Free Class Incremental Learning Method for Environmental Sound Classification
- 通过声学特征变换对齐新旧类别时序特征，构建压缩特征空间。
- 在两个数据集上实现3.7%~3.9%的准确率提升，优于基线模型。
- 无需存储历史样本，适合隐私敏感场景，适用于持续学习任务。
声音蕴含丰富信息，环境声音分类（ESC）在稀有野生动物检测等应用中至关重要。然而世界不断变化，要求ESC模型定期适应新声音。主要挑战是灾难性遗忘——学习新声音时会丢失旧声音的识别能力。现有方法多依赖回放机制，但在数据隐私敏感场景下不适用。无样本方法虽避免了数据存储，但常扭曲旧特征，导致性能下降。为此，我们提出声学特征变换（AFT）技术，将旧类别的时序特征对齐至新空间，包含选择性压缩的特征空间。AFT在不保留历史数据的前提下缓解旧知识遗忘。我们在两个数据集上进行实验，结果表明相比基线模型，准确率提升3.7%至3.9%。
原文摘要 · Abstract (English)
As sounds carry rich information, environmental sound classification (ESC) is crucial for numerous applications such as rare wild animals detection. However, our world constantly changes, asking ESC models to adapt to new sounds periodically. The major challenge here is catastrophic forgetting, where models lose the ability to recognize old sounds when learning new ones. Many methods address this using replay-based continual learning. This could be impractical in scenarios such as data privacy concerns. Exemplar-free methods are commonly used but can distort old features, leading to worse performance. To overcome such limitations, we propose an Acoustic Feature Transformation (AFT) technique that aligns the temporal features of old classes to the new space, including a selectively compressed feature space. AFT mitigates the forgetting of old knowledge without retaining past data. We conducted experiments on two datasets, showing consistent improvements over baseline models with accuracy gains of 3.7\% to 3.9\%.
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。