Machine Learning for Modern Artiﬁcial Intelligence

(1)

Machine Learning for Modern Artificial Intelligence

Hsuan-Tien Lin 林軒田

Dept. of Computer Science and Information Enginnering, National Taiwan University

國立臺灣大學資訊工程學系 December 17, 2020

Keynote talk for International Computer Symposium 2020 &

教育部人工智慧技術及應用人才培育計畫成果發表會

(2)

About Me

Professor

National Taiwan University

Chief Data Science Consultant (former Chief Data Scientist)

Appier Inc.

Co-author Learning from Data

Instructor

NTU-Coursera MOOCs ML Foundations/Techniques

(3)

ML for (Modern) AI

Outline

ML for (Modern) AI

ML Research for Modern AI

ML for Future AI

(4)

From Intelligence to Artificial Intelligence

intelligence: thinking and actingsmartly

• humanly

• rationally

artificialintelligence:computersthinking and actingsmartly

• humanly

• rationally

humanly≈smartly≈rationally

—are humans rational? :-)

(5)

ML for (Modern) AI

Humanly versus Rationally

What if your self-driving car decides one death is better than two—and that one is you? (The Washington Post http://wpo.st/ZK-51)

You’re humming along in your self-driving car, chatting on your iPhone 37 while the machine navigates on its own. Then a swarm of people appears in the street, right in the path of the oncoming vehicle.

Car ActingHumanly tosave my (and passengers’) life, stay on track

Car ActingRationally avoid the crowd and crash the owner forminimum total loss

which issmarter?

—depending on where I am, maybe? :-)

(6)

(Traditional) Artificial Intelligence

Thinking Humanly

• cognitive modeling

—now closer to Psychology than AI

Acting Humanly

• dialog systems

• humanoid robots

• computer vision

Thinking Rationally

• formal logic—now closer to Theoreticians than AI practitioners

Acting Rationally

• recommendation systems

• cleaning robots

• cross-device ad placement

actinghumanly or rationally:

more academia/industry attentions nowadays

(7)

ML for (Modern) AI

Traditional vs. Modern [My] Definition of AI

Traditional Definition

humanly ≈ intelligently ≈ rationally

My Definition

intelligently ≈ easily

is your smart phone ‘smart’? :-)

modern artificial intelligence

=applicationintelligence

(8)

Examples of Application Intelligence

Siri

By Bernard Goldbach [CC BY 2.0]

Amazon Recommendations

By Kelly Sims [CC BY 2.0]

iRobot

By Yuan-Chou Lo [CC BY-NC-ND 2.0]

Vivino

From nordic.businessinsider.com

(9)

ML for (Modern) AI

AI Milestones

logic inference

expert system

machine learning +deep learning

begin 1st winter 2nd winter revolution

1956 1980 1993 2012 time

heat

AI history

• first AI winter: AI cannot solve ‘combinatorial explosion’ problems

• second AI winter: expert system failed to scale

reason of winters: expectation mismatch

(10)

What’s Different Now?

More Data

• cheaper storage

• Internet companies

Faster Computation

• cloud computing

• GPU computing

Better Algorithms

• decades of research

• e.g. deep learning

Healthier Mindset

• reasonable wishes

• key breakthroughs

data-enabledAI: mainstream nowadays

(11)

ML for (Modern) AI

Bigger Data Towards Easier-to-use AI

By deepanker70 on https://pixabay.com/

past

best route by shortest path

present

best route by current traffic

future

best route by predicted travel time

big datacanmake machine look smarter

(12)

Machine Learning Connects Big Data and AI

From Big Data to Artificial Intelligence

big data

ML

artificial intelligence

ingredient tools/steps dish

Photos Licensed under CC BY 2.0 from Andrea Goh on Flickr

“cooking” needs many possible tools & procedures

(13)

ML for (Modern) AI

ML for Modern AI

big data

ML AI

human learning/

analysis

domain knowledge

(HumanI)

method

model expert system

• human sometimesfaster learneroninitial (smaller) data

• industry: black plum is as sweet as white

often important to leverage human learning, especiallyin the beginning

(14)

Application: Tropical Cyclone Intensity Estimation

meteorologists can ‘feel’ & estimate TC intensity from image

TC images

ML

_estimation^intensity

human learning/

analysis

domain knowledge

(HumanI)

CNN polar

rotation invariance

current weather

system

better than current system & ‘trial-ready’

(Chen et al., KDD ’18; Chen et al., Weather & Forecasting ’19)

(15)

Outline

ML for (Modern) AI

ML for Future AI

(16)

Cost-Sensitive Multiclass Classification

(17)

What is the Status of the Patient?

?

By DataBase Center for Life Science;

licensed under CC BY 4.0 via Wikimedia Commons

COVID19 cold healthy

Pictures Licensed under CC BY-SA 3.0 from 1RadicalOne on Wikimedia Commons

• aclassificationproblem

—grouping ‘patients’ into different ‘status’

are all mis-prediction costs equal?

(18)

Patient Status Prediction

error measure = society cost

XXXX

XXXXXX actual

predicted

COVID19 cold healthy

COVID19 0 1000 100000

cold 100 0 3000

healthy 100 30 0

• COVID19 mis-predicted as healthy: very high cost

• cold mis-predicted as healthy: high cost

• cold correctly predicted as cold: no cost

human doctors consider costs of decision;

how about computer-aided diagnosis?

(19)

Our Works

binary multiclass

regular well-studied well-studied

cost-sensitive known(Zadrozny et al., 2003) ongoing(our works, among others)

selected works of ours

• cost-sensitive SVM(Tu and Lin, ICML 2010)

• cost-sensitive one-versus-one(Lin, ACML 2014)

• cost-sensitive deep learning(Chung et al., IJCAI 2016)

why are peoplenot

using thosecool ML works for their AI? :-)

(20)

Issue 1: Where Do Costs Come From?

A Real Medical Application: Classifying Bacteria

• by human doctors: different treatments⇐⇒ serious costs

• cost matrix averaged from two doctors:

Ab Ecoli HI KP LM Nm Psa Spn Sa GBS

Ab 0 1 10 7 9 9 5 8 9 1

Ecoli 3 0 10 8 10 10 5 10 10 2

HI 10 10 0 3 2 2 10 1 2 10

KP 7 7 3 0 4 4 6 3 3 8

LM 8 8 2 4 0 5 8 2 1 8

Nm 3 10 9 8 6 0 8 3 6 7

Psa 7 8 10 9 9 7 0 8 9 5

Spn 6 10 7 7 4 4 9 0 4 7

Sa 7 10 6 5 1 3 9 2 0 7

GBS 2 5 10 9 8 6 5 6 8 0

issue 2: is cost-sensitive classification really useful?

(21)

Cost-Sensitive vs. Traditional on Bacteria Data

. . . . . .

Are cost-sensitive algorithms great?

RBF kernel

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

OVOSVM

csOSRSVM csOVOSVM csFTSVM

algorithms

cost

.

...Cost-sensitive algorithms perform better than regular algorithm

Jan et al. (Academic Sinica) Cost-Sensitive Classiﬁcation on SERS October 31, 2011 15 / 19

(Jan et al., BIBM 2011)

cost-sensitivebetter thantraditional;

but why are peoplestill not

using those cool ML works for their AI? :-)

(22)

Issue 3: Error Rate of Cost-Sensitive Classifiers

The Problem

0.1 0.15 0.2 0.25 0.3

0 0.05 0.1 0.15 0.2

Error (%)

Cost

• cost-sensitive classifier: low cost but high error rate

• traditional classifier: low error rate but high cost

• how can we get theblueclassifiers?: low error rate and low cost

cost-and-error-sensitive:

more suitable forreal-world medical needs

(23)

Improved Classifier for Both Cost and Error

(Jan et al., KDD 2012)

Cost

iris ≈

wine ≈

glass ≈

vehicle ≈

vowel

segment

dna

satimage ≈

usps

zoo

splice ≈

ecoli ≈

soybean ≈

Error

iris

wine

glass

vehicle

vowel

segment

dna

satimage

usps

zoo

splice

ecoli

soybean

now,are people using those cool ML works for their AI? :-)

(24)

Lessons Learned from

Research on Cost-Sensitive Multiclass Classification

? H7N9-infected cold-infected healthy

See Page 16 of the Slides for Sources of the Pictures

1 more realistic (generic) in academia

6=more realistic (feasible) in application e.g. the ‘cost’ ofinputting a cost matrix? :-)

2 cross-domain collaborationimportant

e.g. getting the ‘cost matrix’ fromdomain experts

3 not easy to winhuman trust

—humans are somewhatmulti-objective

(25)

Active Learning by Learning

(26)

Active Learning: Learning by ‘Asking’

labeling isexpensive:

active learning ‘question asking’

—query y_nofchosenx_n

unknown target function f : X → Y

labeled training examples ( , +1), ( , +1), ( , +1)

( , -1), ( , -1), ( , -1)

learning algorithm

A

final hypothesis g≈f

+1

active: improve hypothesis with fewer labels (hopefully) by asking questionsstrategically

(27)

Pool-Based Active Learning Problem

Given

• labeled poolD_l =n

(featurex_n ,label y_n(e.g. IsApple?))oN n=1

• unlabeled pool D_u= nx˜_soS

s=1

Goal

design an algorithm that iteratively

1 strategically querysome˜x_s to get associatedy˜_s

2 move (˜x_s,y˜_s)fromD_utoD_l

3 learnclassifier g^(t)fromD_l

and improvetest accuracy of g^(t) w.r.t#queries

how toquery strategically?

(28)

How to Query Strategically?

Strategy 1

askmost confused question

Strategy 2

askmost frequent question

Strategy 3

askmost debateful question

• choosingone single strategy isnon-trivial:

0 10 20 30 40 50 60

0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8

% of unlabelled data

Accuracy

RAND UNCERTAIN PSDS QUIRE

0 10 20 30 40 50 60

0.4 0.5 0.6 0.7 0.8 0.9

Accuracy

0 10 20 30 40 50 60

0.5 0.6 0.7 0.8 0.9 1

Accuracy

application intelligence: how to choose strategy smartly?

(29)

Idea: Trial-and-Reward Like Human

when do humanstrial-and-reward?

gambling

K strategies:

A₁, A₂, · · · , A_K

tryone strategy

“goodness” of strategy asreward

K bandit machines:

B₁, B2, · · · , BK

tryone bandit machine

“luckiness” of machine asreward

intelligent choice of strategy

=⇒intelligent choice ofbandit machine

(30)

Active Learning by Learning

(Hsu and Lin, AAAI 2015)

K strategies:

A₁, A₂, · · · , A_K

try one strategy

“goodness” of strategy as reward

Given: K existing active learning strategies for t = 1, 2, . . . , T

1 let some bandit modeldecide strategy A_k to try

2 query the ˜x_ssuggested by A_k, and compute g^(t)

3 evaluategoodness of g^(t) asrewardoftrialto update model

proposed Active Learning by Learning (ALBL):

motivated but unrigorousreward design

(31)

Comparison with Single Strategies

UNCERTAINBest

510 15 20 25 30 35 40 45 50 55 60 0.55

0.6 0.65 0.7 0.75 0.8 0.85 0.9

Accuracy ALBL

vehicle

PSDSBest

510 15 20 25 30 35 40 45 50 55 60 0.5

0.55 0.6 0.65 0.7 0.75 0.8

Accuracy ALBL

sonar

QUIREBest

510 15 20 25 30 35 40 45 50 55 60 0.5

0.55 0.6 0.65 0.7 0.75

Accuracy ALBL

diabetes

• no single best strategyfor every data set

—choosing needed

• proposedALBLconsistentlymatches the best

—similar findings across other data sets

‘application intelligence’ outcome:

open-source toolreleased

(https://github.com/ntucllab/libact)

(32)

Have We Made Active Learning More Realistic? (1/2)

Yes!

open-source toollibactdeveloped (Yang, 2017) https://github.com/ntucllab/libact

• including uncertainty, QUIRE, PSDS, . . .,and ALBL

• received>500starsand continuousissues

“libact is a Python package designed tomake ac- tive learning easierfor real-world users”

(33)

Have We Made Active Learning More Realistic? (2/2)

No!

• single-most raisedissue: hard to install on Windows/Mac

—because several strategies requires some C packages

• performance in a recent industry project:

• uncertaintysamplingoften suffices

• ALBLdragged down by bad strategy

“libact is a Python package designed to make active learning easierfor real-world users”

(34)

Lessons Learned from

Research on Active Learning by Learning

by DFID - UK Department for International Development;

licensed under CC BY-SA 2.0 via Wikimedia Commons

1 scalability bottleneckof ‘application intelligence’:

choiceof methods/models/parameter/. . .

2 think outside of themathbox:

‘unrigorous’ usage may begood enough

3 important to bebraveyetpatient

• idea: 2012

• paper:(Hsu and Lin, AAAI 2015); software:(Yang et al., 2017) 4 easy-to-use in design 6=easy-to-use in reality

(35)

Tropical Cyclone Intensity Estimation

(36)

Experienced Meteorologists Can ‘Feel’ and Estimate Tropical Cyclone Intensity from Image

Can ML do the same/better?

• lack ofML-ready datasets

• lack ofmodel that properly utilizes domain knowledge issues addressed in our latest works

(Chen et al., KDD ’18; Chen et al., Weather & Forecasting ’19)

(37)

Recall: Flow behind Our Proposed Model

TC images

ML

_estimation^intensity

human learning/

analysis

domain knowledge

(HI)

CNN polar

rotation invariance

current weather

system

is proposedCNN-TCbetter than current weather system?

(38)

Results

RMS Error

ADT 11.75

AMSU 14.40

SATCON 9.66 CNN-TC 9.03

CNN-TC much betterthan current weather system (SATCON)

why are peoplenot using thiscool ML model? :-)

(39)

Lessons Learned from

Research on Tropical Cyclone Intensity Estimation

1 again,cross-domain collaborationimportant e.g. even from ‘organizing data’ to be ML-ready

2 not easy to claimproduction ready

—can ML be used for ‘unseenly-strongTC’?

3 good AI system requiresboth human and machine learning

—still an ‘art’ to blend the two

(40)

Outline

ML for (Modern) AI

ML for Future AI

(41)

ML for Future AI

AI: Now and Next

2010–2015: AI | AI becomes promising, e.g.

• initial success of deep learningon ImageNet

• mature tools for SVM (LIBSVM) and others

2016–2020: AI + AI becomes competitive, e.g.

• super-human performance of alphaGoand others

• all big technology companies becomeAI-first

2021–: AI × AI becomes necessary

• “You’ll not be replaced by AI, butby humans who know how to use AI”

(Sun, Chief AI Scientist of Appier, 2018)

(42)

Needs of ML for Future AI

more creative win humanrespect

e.g. Appier’s 2018 work on

design matching clothes

(Shih et al., AAAI 2018)

more explainable win humantrust

e.g. my students’

work on

automatic bridge bidding

(Yeh et al., IEE ToG 2018)

more interactive win humanheart

e.g. my student’s work (w/ DeepQ) on efficient disease diagonsis

(Peng et al., NeurIPS 2018)

(43)

ML for Future AI

Summary

• ML for (Modern) AI:

tools + human knowledge ⇒easy-to-use application

• ML Research for Modern AI:

need to bemore open-minded

—in methodology, in collaboration, in KPI

• ML for Future AI:

crucial to be ‘human-centric’

Thank you! Questions?