arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.00816eess.AScs.AIcs.LGcs.SDeess.SP
收藏

用于自监督表示学习的对比分离编码

Contrastive Separative Coding for Self-supervised Representation Learning

Tencent AI Lab(腾讯人工智能实验室)

Jun Wang, Max W. Y. Lam, Dan Su, Dong Yu

摘要:

为了从语音数据的长序列建模中提取鲁棒的深度表示,我们提出了一种自监督学习方法,即对比分离编码(CSC)。我们的关键发现是通过将目标信号与对比干扰信号分离来学习这种表示。首先,构建了一个多任务分离编码器来提取共享的可分离和可区分的嵌入;其次,我们提出了一个强大的交叉注意机制,在各种干扰条件下对说话人表示执行,允许模型关注并全局聚合最关键的信息来回答“查询”(当前自下而上嵌入),同时较少关注干扰、噪声或无关部分;最后,我们构造了一种新的概率对比损失,它估计并最大化了表示与全局说话人向量之间的互信息。虽然大多数先前的无监督方法都专注于预测未来,相邻或缺失的样本,但我们对预测干扰样本采取了不同的观点。此外,我们的对比分离损失是免费的负采样。实验表明,我们的方法可以学习有用的表示,在不利条件下实现强大的说话者验证性能。

英文摘要:

To extract robust deep representations from long sequential modeling of speech data, we propose a self-supervised learning approach, namely Contrastive Separative Coding (CSC). Our key finding is to learn such representations by separating the target signal from contrastive interfering signals. First, a multi-task separative encoder is built to extract shared separable and discriminative embedding; secondly, we propose a powerful cross-attention mechanism performed over speaker representations across various interfering conditions, allowing the model to focus on and globally aggregate the most critical information to answer the "query" (current bottom-up embedding) while paying less attention to interfering, noisy, or irrelevant parts; lastly, we form a new probabilistic contrastive loss which estimates and maximizes the mutual information between the representations and the global speaker vector. While most prior unsupervised methods have focused on predicting the future, neighboring, or missing samples, we take a different perspective of predicting the interfered samples. Moreover, our contrastive separative loss is free from negative sampling. The experiment demonstrates that our approach can learn useful representations achieving a strong speaker verification performance in adverse conditions.

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。

↑