用纯注意力机制实现轻量语音质量评估，更准更实用。
AttentiveMOS: A Lightweight Attention-Only Model for Speech Quality Prediction
- 仅用Transformer注意力层提取局部与全局语音特征
- 在三个数据集上优于基线模型，尤其适应噪声标签
- 轻量化设计适合真实场景部署，无需大预训练模型
针对无参考语音质量评估的研究已发展出直接基于语音波形预测平均意见分（MOS）的模型。这些模型通常依赖卷积层提取局部特征，并使用过大预训练网络的嵌入来增强泛化能力。本文提出一种基于Swin Transformer和标准Transformer层的纯注意力模型，分别用于提取局部上下文特征和全局话语特征。自注意力机制擅长处理序列数据，其轻量化设计在有限MOS数据集上提升泛化性能，增强实际应用性。我们采用序列式自教学策略训练网络，以改善受听者评分噪声影响的MOS标签泛化能力。在三个数据集上的实验验证了该设计的有效性，结果优于基线模型。
原文摘要 · Abstract (English)
Research in modeling subjective metrics for quality assessment has led to the development of no-reference speech models that directly operate on utterance waveforms to predict the Mean Opinion Score (MOS). These models often rely on convolutional layers for local feature extraction and embeddings from impractically large pretrained networks to enhance generalization. We propose an attention-only model based on Swin transformer and standard transformer layers to extract local context features and global utterance features, respectively. The self-attention operator excels at processing sequences, and our lightweight design enhances generalization on limited MOS datasets while improving real-world applicability. We train our network using a sequential self-teaching strategy to improve generalization on MOS labels affected by noise in listener ratings. Experiments on three datasets confirm the effectiveness of our design and demonstrate improvement over baseline models.
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。