- Lecture 1.1 - What is ML 什麼是機器學習
- Lecture 1.2 - What is DL 什麼是深度學習
- Lecture 1.3 - How to Apply 如何應用深度學習
- Lecture 2.1 - How to Train a Model 如何訓練模型
- Lecture 2.2 - What is Model 什麼是模型
- Lecture 2.3 - What does the Good Function Mean 什麼叫做好的Function呢
- Lecture 2.4 - How can we Pick the Best Function 如何找出最好的Function
- Lecture 2.5 - Backpropagation 效率地計算大量參數
- TA Recitation - Optimization
- Lecture 3.1 - Word Representations
- Lecture 3.2 - Language Modeling 語言模型
- Lecture 3.3 - Recurrent Neural Network 詳細解析
- Lecture 3.4 - RNN Applications RNN各式應用
- TA Recitation - Practical Tips
- Lecture 4.1 - Attention Mechanism 注意力機制
- Lecture 4.2 - Attention Applications 注意力的各式應用
- Assignment 1 Tutorial
- Lecture 5.1 - Word Representation Review 詞向量各式表示法
- Lecture 5.2 - Word2Vec 詞向量
- Lecture 5.3 - Word2Vec Training 如何訓練詞向量
- Lecture 5.4 - Negative Sampling
- Lecture 5.5 - Word2Vec Variants 各種訓練的變化方式
- Lecture 5.5 - Word2Vec Variants 各種訓練的變化方式
- Lecture 5.6 - GloVe 詞向量
- Lecture 5.7 - Word Vector Evaluation 如何評估詞向量的好壞
- Lecture 5.8 - Contextualized Word Embeddings 前後文相關之詞向量
- Lecture 5.9 - ELMo 芝麻街家族之起源
- Lecture 6.1 - Basic Attention 基本注意力模型複習
- Lecture 6.2 - Self Attention 新注意力機制
- Lecture 6.3 - Multi-Head Attention
- Lecture 6.4- Transformer
- Lecture 6.5- BERT 進擊的芝麻街巨人
- TA Recitation- More on Embeddings
- Lecture 7.1 - Transformer-XL 處理超長輸入的變形金剛
- Lecture 7.2 - XLNet 兼顧AR及AE好處的模型
- Lecture 7.3 - RoBERTa SpanBERT XLM 簡單有用的改進方法
- Lecture 7.4- ALBERT 如何讓BERT縮小卻依然好用呢
- TA Recitation - More on Transformers
- Lecture 8.1- Deep Reinforcement Learning Introduction
- Lecture 8.2- Markov Decision Process
- Lecture 8.3- Reinforcement Learning
- Lecture 8.4- Value-Based RL Approach
- Lecture 8.5- Advanced DQN
- Lecture 9.1- Policy Gradient
- Lecture 9.2- Actor Critic
- Lecture 10.1- Natural Language Generation
- Lecture 10.2- Decoding Algorithm
- Lecture 10.3- NLG Evaluation
- Lecture 10.4- RL for NLG (20-05-12)
- TA Recitation- RL for Dialogues (20-05-12)
- GAN (Quick Review)
- GAN Lecture 4 (2018)- Basic Theory
- GAN Lecture 6 (2018)- WGAN EBGAN
- Lecture 11.1- Unsupervised Learning Introduction
- Lecture 11.2- Autoencoder & Variational Autoencoder
- Lecture 11.3- Distant Supervision & Multi-Task Learning
- Lecture 12.1- Conversational AI Introduction 對話AI簡介
- Lecture 12.2- Task-Oriented Dialogues 任務型對話
- Lecture 12.3- Chit-Chat Social Bots 聊天型對話
- Lecture 13.1- Robustness 對話系統的強健性 (20-06-16)
- Lecture 13.2- Scalability 對話系統的擴展性
- Final Project- Rules & Grading
本课程主要讲解如何利用深度学习算法来解决各种实际应用场景问题,学生学习如何使用这些深度学习算法,以及为什么要使用这些算法。本课程希望学生在课堂上学习理论,并通过做作业和最后的项目来学习实施方法。 注意:如果已修过类似的课程,例如,李宏毅老师的课程,则无需修此课程。
课程涵盖了深度学习和表示学习中的最新技术,重点包括监督/自监督学习、嵌入方法、度量学习、卷积网络和循环网络,并应用于计算机视觉、自然语言理解和语音识别。
This course is enable students to learn how and why to apply deep learning to tackle various practical problems, where the students are expected to learn the theory during the class and learn the implementation by doing assignments and final projects.
Lecture 0 2019/02/19 Course Logistics [slides]
Registration: [Google Form]
Lecture 1 2019/02/26 Introduction [slides] (video)
Guest Lecture (R103) [PyTorch Tutorial]
Lecture 2 2019/03/05 Neural Network Basics [slides] (video)
Suggested Readings:
[Linear Algebra]
[Linear Algebra Slides]
[Linear Algebra Quick Review]
A1 2019/03/05 A1: Dialogue Response Selection [A1 pages]
Lecture 3 2019/03/12 Backpropagation [slides] (video)
Word Representation [slides] (video)
Suggested Readings:
[Learning Representations]
[Vector Space Models of Semantics]
[RNNLM: Recurrent Neural Nnetwork Language Model]
[Extensions of RNNLM]
[Optimzation]
Lecture 4 2019/03/19 Recurrent Neural Network [slides] (video)
Basic Attention [slides] (video)
Suggested Readings:
[RNN for Language Understanding]
[RNN for Joint Language Understanding]
[Sequence-to-Sequence Learning]
[Neural Conversational Model]
[Neural Machine Translation with Attention]
[Summarization with Attention]
[Normalization]
A2 2019/03/19 A2: Contextual Embeddings [A2 pages]
Lecture 5 2019/03/26 Word Embeddings [slides] (video)
Contextual Embeddings - ELMo [slides] (video)
Suggested Readings:
[Estimation of Word Representations in Vector Space]
[GloVe: Global Vectors for Word Representation]
[Sequence Tagging with BiLM]
[Learned in Translation: Contextualized Word Vectors]
[ELMo: Embeddings from Language Models]
[More Embeddings]
2019/04/02 Spring Break A1 Due
Lecture 6 2019/04/09 Transformer [slides] (video)
Contextual Embeddings - BERT [slides] (video)
Gating Mechanism [slides] (video)
Suggested readings:
[Contextual Word Representations Introduction]
[Attention is all you need]
[BERT: Pre-training of Bidirectional Transformers]
[GPT: Improving Understanding by Unsupervised Learning]
[Long Short-Term Memory]
[Gated Recurrent Unit]
[More Transformer]
Lecture 7 2019/04/16 Reinforcement Learning Intro [slides] (video)
Basic Q-Learning [slides] (video)
Suggested Readings:
[Reinforcement Learning Intro]
[Stephane Ross' thesis]
[Playing Atari with Deep Reinforcement Learning]
[Deep Reinforcement Learning with Double Q-learning]
[Dueling Network Architectures for Deep Reinforcement Learning]
A3 2019/04/16 A3: RL for Game Playing [A3 pages]
Lecture 8 2019/04/23 Policy Gradient [slides] (video)
Actor-Critic (video)
More about RL [slides] (video) Suggested Readings:
[Asynchronous Methods for Deep Reinforcement Learning]
[Deterministic Policy Gradient Algorithms]
[Continuous Control with Deep Reinforcement Learning]
A2 Due
Lecture 9 2019/04/30 Generative Adversarial Networks [slides] (video)
(Lectured by Prof. Hung-Yi Lee)
Lecture 10 2019/05/07 Convolutional Neural Networks [slides]
A4 2019/05/07 A4: Drawing [A4 pages]
2019/05/14 Break A3 Due
Lecture 11 2019/05/21 Unsupervised Learning [slides]
NLP Examples [slides]
Project Plan [slides]
Special 2019/05/28 Company Workshop Registration: [Google Form]
2019/06/04 Break A4 Due
Lecture 12 2019/06/11 Project Progress Presentation
Course and Career Discussion
Special 2019/06/18 Company Workshop Registration: [Google Form]
Lecture 13 2019/06/25 Final Presentation