Applied Deep Learning 2021

Homework of Applied Deep Learning 2021 at NTU (Lectured by Yun-Nung / Vivian, Chen)

This course is about Natural Language Processing, including Sentiment or Intent Classification, Slot Tagging, Question Answering, Text Generation, and so on.

Homework 0 - Sentiment Classification

Task description (Slides)

A binary classification problem for sentiment analysis in Chinese

Source Code

• https://github.com/joe0123/ADL2021/tree/master/HW0

Inclass Competition

Kaggle for Sentiment Analysis

• Public Leaderboard: Rank 1/90
• Private Leaderboead: Rank 1/90

Homework 1 - Intent Classification and Slot Tagging

Task description (Slides)

There are two problems in this homework. One is multi-class intent classification, and the other is BIO slot tagging.

Source Code

• For RNN-based method: https://github.com/joe0123/ADL2021/tree/master/HW1
• For BERT-based method: https://github.com/joe0123/ADL2021/tree/master/HW1_BERT

Inclass Competition (Only RNN-based method is allowed)

Kaggle for Intent Classification

• Public Leaderboard: Rank 6/91
• Private Leaderboead: Rank 4/91

Kaggle for Slot Tagging

• Public Leaderboard: Rank 10/89
• Private Leaderboead: Rank 25/89

Homework 2 - Question Answering

Task description (Slides)

A Question Answering problem in Chinese, which can be decomposed into two subtasks (context selection + span selection).

Source Code

• https://github.com/joe0123/ADL2021/tree/master/HW2

Inclass Competition

• Public Leaderboard: Rank 9/76
• Private Leaderboead: Rank 9/76

Homework 3 - Natural Language Generation

Task description (Slides)

A Natural Language Generation (NLG) problem, entitling news in Chinese.

Source Code

• https://github.com/joe0123/ADL2021/tree/master/HW3