Completed
How Much Transcribed Audio Do We Need?
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Multilingual Representations for Low-Resource Speech Processing
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 Why Care About Low-Resource Speech Processing?
- 3 How Much Transcribed Audio Do We Need?
- 4 Why Do We Need All That Training Data?
- 5 Multilingual Features
- 6 The IARPA Babel Program
- 7 Babel Languages
- 8 Limited resources
- 9 What is keyword search, and why focus on it?
- 10 How do we measure keyword search performance?
- 11 Properties of term-weighted value
- 12 Take-Home Messages
- 13 Three Ways of Looking at Speech
- 14 Deep Neural Network
- 15 A Stacked DNN Architecture
- 16 Convolutional Neural Network
- 17 Considered 2 CNN Architectures
- 18 Recurrent Neural Network
- 19 Bidirectional LSTM Architecture
- 20 Three Use Cases
- 21 More Expressive Architectures Make a Big Difference
- 22 Fixed Features Allow for Rapid Development
- 23 Our partners
- 24 Babel resources
