awesome-speech

curated paper, notes, codes storage for speech recognition,speech synthesis, signal process...

Challenge for accuracy

• challenge https://www.kaggle.com/c/tensorflow-speech-recognition-challenge#
• analysis https://dinantdatascientist.blogspot.com/2018/02/kaggle-tensorflow-speech-recognition.html

SOTA https://paperswithcode.com/task/speech-recognition

Papers

Deep Speech: Scaling up end-to-end speech recognition https://arxiv.org/pdf/1412.5567v2.pdf

• noise environment

Deep Speech 2: End-to-End Speech Recognition in English and Mandarin https://arxiv.org/pdf/1512.02595v1.pdf

• integrated in tensorflow model

SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition https://arxiv.org/pdf/1904.08779v3.pdf

English Conversational Telephone Speech Recognition by Humans and Machines https://arxiv.org/pdf/1703.02136v1.pdf

STATE-OF-THE-ART SPEECH RECOGNITION USING MULTI-STREAM SELF-ATTENTION WITH DILATED 1D CONVOLUTIONS https://arxiv.org/pdf/1910.00716v1.pdf

Purely sequence-trained neural networks for ASR based on lattice-free MMI https://www.danielpovey.com/files/2016_interspeech_mmi.pdf

Kaldi recipe

• kaldi asr recipes with languages examples https://github.com/kaldi-asr/kaldi/tree/master/egs

• korean recipe (zeroth) https://github.com/kaldi-asr/kaldi/tree/master/egs/zeroth_korean/s5

한국어 preprocess

http://speech.cbnu.ac.kr/srhome/technology/korean_recog.html