探索用于说话人验证的二分类损失
Exploring Binary Classification Loss For Speaker Verification
AI总结：
针对说话人验证中闭集训练与开集测试失配问题，提出基于二分类器成对训练的 SphereFace2 框架，在 Voxceleb 上优于现有损失，并具噪声标签鲁棒性。
摘要：
闭集训练与开集测试之间的失配通常会导致说话人验证任务性能显著下降。对于现有损失函数，基于度量学习的目标强烈依赖于搜索有效样本对，这可能阻碍进一步改进；而流行的多分类方法在未见说话人上评估时通常会出现性能退化。本文引入 SphereFace2 框架，使用若干二分类器以成对方式训练说话人模型，而非执行多分类。得益于这种学习范式，它能够有效缓解训练与评估之间的差距。在 Voxceleb 上进行的实验表明，SphereFace2 优于其他现有损失函数，尤其在困难试验上表现突出。此外，大间隔微调策略被证明与其兼容，可用于进一步提升性能。最后，SphereFace2 还表现出对类别级噪声标签的强鲁棒性，有潜力应用于伪标签估计不准确的半监督训练场景。代码可在 https://github.com/Hunterhuan/sphereface2_speaker_verification 获取。
英文摘要：
The mismatch between close-set training and open-set testing usually leads to significant performance degradation for speaker verification task. For existing loss functions, metric learning-based objectives depend strongly on searching effective pairs which might hinder further improvements. And popular multi-classification methods are usually observed with degradation when evaluated on unseen speakers. In this work, we introduce SphereFace2 framework which uses several binary classifiers to train the speaker model in a pair-wise manner instead of performing multi-classification. Benefiting from this learning paradigm, it can efficiently alleviate the gap between training and evaluation. Experiments conducted on Voxceleb show that the SphereFace2 outperforms other existing loss functions, especially on hard trials. Besides, large margin fine-tuning strategy is proven to be compatible with it for further improvements. Finally, SphereFace2 also shows its strong robustness to class-wise noisy labels which has the potential to be applied in the semi-supervised training scenario with inaccurate estimated pseudo labels. Codes are available in https://github.com/Hunterhuan/sphereface2_speaker_verification
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。
