李宏毅(李弘毅)分別于2010年和2012年獲得國(guó)立臺(tái)灣大學(xué)碩士和博士學(xué)位。2012年9月至2013年8月，Academia Sinica信息技術(shù)創(chuàng)新研究中心博士后。2013年9月至2014年7月，麻省理工學(xué)院計(jì)算機(jī)科學(xué)和人工智能實(shí)驗(yàn)室( CSAIL )語(yǔ)言系統(tǒng)組的訪問(wèn)科學(xué)家。現(xiàn)為國(guó)立臺(tái)灣大學(xué)電氣工程系助理教授，并獲大學(xué)計(jì)算機(jī)科學(xué)與信息工程系聯(lián)合委任。他的研究側(cè)重于機(jī)器學(xué)習(xí)(尤其是深度學(xué)習(xí))、語(yǔ)言理解和語(yǔ)音識(shí)別。

Machine Learning 

2020

李宏毅

Hung-yi Lee

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

本學(xué)期總共有十五個(gè)作業(yè) (每項(xiàng)作業(yè)滿分皆為10 分，

學(xué)期成績(jī)以分?jǐn)?shù)最高的前十個(gè)作業(yè)計(jì)算)

f ( ) = f ( ) = f ( ) = f ( ) =

機(jī)器學(xué)習(xí)就是自動(dòng)找函式

• Speech Recognition

• Image Recognition

• Playing Go

• Dialogue System

“Cat”

“How are you”

“5-5”

“How are you?” “I am fine.”

(what the user said) (system response)

(next move)

你想找什麼樣的函式？

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

f

PM2.5 today

PM2.5 yesterday

…….

PM2.5 tomorrow

(scalar)

The output of the 

function is a scalar.

Regression

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

Input f Yes or No

(sentence) (pos or neg)

Binary 

Classification 

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

f

Input Class 1, Class 2, … Class N

Multi-class 

Classification 

麵包 蛋 湯

Generation (生成)

產(chǎn)生有結(jié)構(gòu)的複雜東西

(例如：文句、圖片)

擬人化的講法—?jiǎng)?chuàng)造

Regression, 

Classification

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

Generation

翻譯：產(chǎn)生文句

產(chǎn)生二次元

人物

怎麼告訴機(jī)器

你想找什麼樣的函式？

Supervised Learning

𝒇 “Cat”

𝑥 𝑦 𝑦1:“Cat” 𝑦2:“Cat”

𝑦3:“Dog” 𝑦4:“Dog”

𝑥1: 𝑥2: 𝑥3: 𝑥4:

Labelled Data

函式的 Loss

𝒇𝟏

“Dog”/“Dog”/

“Dog”/“Dog”

𝑥1 / 𝑥2 / 𝑥3 / 𝑥4

Loss = 50% 

𝑦1:“Cat” 𝑦2:“Cat”

𝑦3:“Dog” 𝑦4:“Dog”

𝑥1: 𝑥2: 𝑥3: 𝑥4:

Labelled Data

函式的 Loss

𝒇𝟐

“Cat”/“Cat”/

“Dog”/“Dog”

𝑥1 / 𝑥2 / 𝑥3 / 𝑥4

Loss = 0% 

接下來(lái)機(jī)器會(huì)自動(dòng)找出

Loss 最低的函式

𝑦1:“Cat” 𝑦2:“Cat”

𝑦3:“Dog” 𝑦4:“Dog”

𝑥1: 𝑥2: 𝑥3: 𝑥4:

Labeled Data

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

Supervised Learning

Reinforcement Learning 

Supervised v.s. Reinforcement

• Supervised:

• Reinforcement Learning

Next move:

“5-5”

Next move:

“3-3”

First move …… many moves …… Win!

Alpha Go is supervised learning + reinforcement learning.

(Reward)

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

Reinforcement Learning

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

Unsupervised 

Learning

What can machine learn 

from unlabeled images? 

機(jī)器怎麼

找出你想要的函式？

Use RNN

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

限制函式尋找範(fàn)圍

Linear

Network Architecture

Use CNN

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

函式尋找方法 – Gradient Descent

Implement the 

algorithm by yourself

Deep Learning Framework 

(3/26 PyTorch 教學(xué)、會(huì)錄影)

前沿研究

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

This is a “cat”

Because …

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

Add

noise

This is a “cat”

Star Fish …

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

This is a “cat”

CNN required

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

This is a “cat”

我不知道

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

99.5% 57.5%

Training 

Data

Testing 

Data

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

Meta Learning = Learn to learn

• Now we design the learning algorithm

• Can machine learn the learning algorithm? 

program 

for learning

I can learn!

program designing 

program 

for learning 

program 

for learning

I can learn!

能不能讓機(jī)器聰明一點(diǎn)？

天資不佳卻勤奮不懈？

http://web.stanford.edu/class/psych209/Readings/LakeEtAlBBS.pdf

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

終身學(xué)習(xí) (Life-long Learning) 

Learning 

Task 1

Learning 

Task 2

Learning 

Task 3

…

I can solve 

task 1.

I can solve 

tasks 1&2.

I can solve 

tasks 1&2&3.

Life-Long Learning (終身學(xué)習(xí)), Continuous Learning, 

Never Ending Learning, Incremental Learning

CNN required

Regression 

Classification

RNN Seq2seq

Meta Learning

Unsupervised 

Learning

(Auto-encoder)

Life-long 

Learning

Reinforcement 

Learning

CNN

Explainable AI

Adversarial 

Attack 

Network 

Compression 

Anomaly 

Detection

GAN

Transfer Learning

(Domain Adversarial 

Learning)

Easy Normal Challenging 

(數(shù)分鐘完成) (數(shù)小時(shí)完成) (數(shù)日完成)

Kaggle

(僅供參考)

Learning order

課程網(wǎng)頁(yè)

• http://speech.ee.ntu.edu.tw/~tlkagk/courses_ML20.html

完全可以在家自學(xué)!

課程網(wǎng)頁(yè)

在寫作業(yè)前先線上學(xué)習(xí)

課程網(wǎng)頁(yè) 所有作業(yè)都有 Colab 範(fàn)例，

照著做就完成一半!

課程網(wǎng)頁(yè) 作業(yè)的要求都在這裡

(錄影預(yù)計(jì) 3/12 全數(shù)完成)

所有作業(yè)皆已經(jīng)公告，現(xiàn)在就可以開始做了

課程網(wǎng)頁(yè) 上課補(bǔ)充的是相關(guān)主題最新的知識(shí)，

和作業(yè)沒(méi)有直接關(guān)連 (會(huì)錄影)

10:20 開始 ，3/26 後每星期都有 (國(guó)定假日除外)

課程網(wǎng)頁(yè)

每一個(gè)作業(yè)都有死線

以後每週四上午 9:10 – 10:00 就是助教時(shí)間

FB 社團(tuán)

• 社團(tuán): “Machine Learning (2020,Spring)”

• https://www.facebook.com/groups/1099602297060276/

歡迎同學(xué)們提問(wèn) ☺

感謝助教群! ! !

助教信箱：

ntu-ml-2020spring-ta@googlegroups.com