Speech Recognition Tutorial

한국어 음성 인식 프로젝트인 KoSpeech 프로젝트를 진행하며 팀원들끼리 스터디한 내용들을 공개한 repository입니다.
부족한 점이 많은 학부생들끼리 공부한 내용이니 잘못 이해한 부분이 있을 수 있습니다.
오류사항, 피드백 등 모두 환영합니다.

Seminar

• Signal Processing

• Bayesian

• Mel Frequency Cepstral Coefficients (MFCC)

• Convolutional Neura Network (CNN)

• Gradient Descent Optimization Algorithms

• Recurrent Neural Network (RNN)

• LSTM & GRU

• Bidirectional LSTM

• Sequence-to-Sequence

• Teacher Forcing

• Attention Mechanism

• Listen, Attend and Spell

• Attention Based Models for Speech Recognition

• Attention Is All You Need

• SpecAugment

• Learning rate Scheduling & Label Smoothing

• Language Model & Fusinging

• Perplexity

• An Analysis of Incorporating an External Language Model Into Sequence-to-Sequence Model

• Ensemble

• Word Embedding

Paper Review

• Listen, Attend and Spell

• Deep Speech

• Attention Based Models for Speech Recognition

• Attention Is All You Need

• AN ANALYSIS OF INCORPORATING AN EXTERNAL LANGUAGE MODEL INTO A SEQUENCE-TO-SEQUENCE MODEL

• STATE-OF-THE-ART SPEECH RECOGNITION WITH SEQUENCE-TO-SEQUENCE MODE

• SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition

• ClovaCall: Korean Goal-Oriented Dialog Speech Corpus for Automatic Speech Recognition of Contact Centers

Author

Soohwan Kim, Seyoung Bae, Cheolhwang Won