Created 07/28/2020 at 10:45PM

Paper

• Audio Word2Vec: Unsupervised Learning of Audio Segment Representations using Sequence-to-sequence Autoencoder
• SpeechBERT: An Audio-and-text Jointly Learned Language Model for End-to-end Spoken Question Answering (李宏毅团队)
• Improving Transformer-based Speech Recognition Using Unsupervised Pre-training (滴滴的语音识别预训练模型)