arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2007.04134eess.AScs.CLcs.CVcs.LGcs.MMcs.SD
收藏

基于联合视听自监督的语音表征学习

Learning Speech Representations from Raw Audio by Joint Audiovisual Self-Supervision

Imperial College London(帝国理工学院);Samsung AI Centre, Cambridge(三星人工智能中心(剑桥));Facebook London(Facebook(伦敦))

Abhinav Shukla, Stavros Petridis, Maja Pantic

摘要:

音频和视觉模态之间的直观交互对于跨模态自监督学习是有价值的。这一概念已被证明为通用的视听任务,如视频动作识别和声学场景分类。然而,自我监督仍然没有充分探讨视听讲话。我们提出了一种从原始音频波形中学习自监督语音表示的方法。我们训练一个原始的音频编码器相结合的音频只有自我监督(通过预测信息音频属性)与视觉自我监督(通过生成说话的脸从音频)。视觉托词任务驱动音频表示来捕获与嘴唇运动相关的信息。这丰富了音频编码器的视觉信息和编码器可以用于评估没有视觉模态。我们的方法相对于现有的自监督音频特征在已建立的孤立词分类基准上具有竞争力的性能,并且在从较少的标签中学习方面显着优于其他方法。值得注意的是,我们的方法还优于完全监督训练,从而为语音相关任务提供了强大的初始化。我们的研究结果表明,在视听语音学习良好的音频表示的多模态自我监督的潜力。

英文摘要:

The intuitive interaction between the audio and visual modalities is valuable for cross-modal self-supervised learning. This concept has been demonstrated for generic audiovisual tasks like video action recognition and acoustic scene classification. However, self-supervision remains under-explored for audiovisual speech. We propose a method to learn self-supervised speech representations from the raw audio waveform. We train a raw audio encoder by combining audio-only self-supervision (by predicting informative audio attributes) with visual self-supervision (by generating talking faces from audio). The visual pretext task drives the audio representations to capture information related to lip movements. This enriches the audio encoder with visual information and the encoder can be used for evaluation without the visual modality. Our method attains competitive performance with respect to existing self-supervised audio features on established isolated word classification benchmarks, and significantly outperforms other methods at learning from fewer labels. Notably, our method also outperforms fully supervised training, thus providing a strong initialization for speech related tasks. Our results demonstrate the potential of multimodal self-supervision in audiovisual speech for learning good audio representations.

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。

↑