→ Last Lesson we had cover on the introduction of the MDD (Mispronunciation Detection and Diagnosis), and distinguish with the ASR task by analysis each task goals and components in each system designing.=

→ In this lesson - we shall focus on how to implement the feature processing - an important step for any Machine Learning model can obtain the knowledge from the raw input, not just the ASR nor MDD system either.

→ This lesson will capture !

1. What is the common input source that the MDD model will accept?
2. What type of feature can be used to train & develop the MDD-System? & How to yield them
3. Additional: Which dataset you can use

Notebook-LM link for TL,DR people & more interactive learning experience: https://notebooklm.google.com/notebook/b02f6aad-3ec6-4299-9df8-07c855e490f7

Podcast audio version:

1. What is the common input source that MDD will accept

2. What type of feature can be used to train & develop the MDD-System? & How to yield them ?

3. Additional - Which Dataset that you can use for MDD task

Reference for lesson 2:

In the next lesson - we shall cover the loss & optimization goal that researcher use in creating training goal in MDD-challenge - Lesson 3: Loss & Optimization goal in MDD-Model Training