面向增强说话人验证的集成式多层次知识蒸馏
Integrated Multi-Level Knowledge Distillation for Enhanced Speaker Verification
AI总结：
针对现有说话人验证知识蒸馏方法未考虑语音多层次时序特性的问题，提出IML-KD方法，通过整合多时长语音的时序上下文与输入敏感表示并进行多层次对齐训练，在VoxCeleb1上使EER降低5%。
摘要：
知识蒸馏（KD）通过将训练完备的大模型（教师模型）的知识迁移到更小、更紧凑的模型（学生模型）中以提升效率和便携性，被广泛应用于说话人验证（SV）等音频任务中。现有的说话人验证知识蒸馏方法多借鉴图像处理领域的方案，聚焦于拟合预测概率与隐层表示，但未考虑语音音频的多层次时序特性。本文提出一种新型知识蒸馏方法——集成式多层次知识蒸馏（IML-KD），用于将语音的多时间尺度特征知识从教师模型迁移到学生模型。在IML-KD中，来自教师模型的时序上下文信息被整合到基于集成梯度的新型输入敏感表示中，这些表示来自不同时长的语音片段；学生模型通过输出端的多层次对齐训练来学习推断这些表示。我们在VoxCeleb1数据集上开展说话人验证实验以评估所提方法，结果表明IML-KD显著提升了知识蒸馏性能，将等错误率（EER）降低了5%。
英文摘要：
Knowledge distillation (KD) is widely used in audio tasks, such as speaker verification (SV), by transferring knowledge from a well-trained large model (the teacher) to a smaller, more compact model (the student) for efficiency and portability. Existing KD methods for SV often mirror those used in image processing, focusing on approximating predicted probabilities and hidden representations. However, these methods fail to account for the multi-level temporal properties of speech audio. In this paper, we propose a novel KD method, i.e., Integrated Multi-level Knowledge Distillation (IML-KD), to transfer knowledge of various temporal-scale features of speech from a teacher model to a student model. In the IML-KD, temporal context information from the teacher model is integrated into novel Integrated Gradient-based input-sensitive representations from speech segments with various durations, and the student model is trained to infer these representations with multi-level alignment for the output. We conduct SV experiments on the VoxCeleb1 dataset to evaluate the proposed method. Experimental results demonstrate that IML-KD significantly enhances KD performance, reducing the Equal Error Rate (EER) by 5%.
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。
