1 of 57

<< Wrapping Up >>

Week 11

CCAI 9012

GenAI Solutions to Global Challenges: �Using AI Creatively and Responsibly

Kam-Ming Mark Tam

Class Coordinator

Impressionist-style scene of human–AI co-existence

Generated with ChatGPT (OpenAI), 2026.

Kanxuan HE

Naixiang GAO

Yifan XIE

Chao TANG

May Loaay Mohamed EL-HADIDI

Joseph Jing Hymn WONG

Christina Ka Man CHU

Charles Wai Lam TO

Shum Nga MAN

BT 1 (ARCH 2056/7330) Building Technology 1 : Building Principles

07/04/2026

1

Kam-Ming Mark TAM

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

1

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

2 of 57

Announcements

Case Study – A.1.1

Quick Topic Strengthening

[…Break…]

Course Project Consultation

​

from 18:00: Crit-Style Peer Review

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

2

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

3 of 57

Announcements

Case Study – A.1.1

Quick Topic Strengthening

[…Break…]

Course Project Consultation

​

from 18:00: Crit-Style Peer Review

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

3

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

4 of 57

Review of Timeline

Extending LLMs

Reinforcing Concepts

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

4

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

5 of 57

Review of Timeline

Case Study Extension

Extending LLMs

Reinforcing Concepts

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

5

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

6 of 57

Review of Timeline

Final Course Project Submission

Extending LLMs

Reinforcing Concepts

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

6

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

7 of 57

Course Project Rubric

Four Components to Deliver

  • i. Annotation & Title
    • 300 words + 200 words * 3 parts

​

  • ii. Key Artefact = Paper
    • Introduction, method, results, conclusion…

​

  • iv. Evidences (Assignment-Specific)

​

  • iii. Vignette (Short-Form Video)
    • 5 to 8 minutes long
    • highlight key insights & significance

i

ii

iii

iv

&

ii

iv

i

iii

+ image slides with explanatory context

Column 3

feel free to incorporate content into video to support

storytelling

Column 1

​

Column 3

​

Column 2

​

Column 2

Column 1

<< Final Course Project Title >>�Group X

<< Final Course Project Title >>�Group X

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

7

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

8 of 57

Let’s open some outline files…

General Course Project Outline

Final Project Submission Outline

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

8

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

9 of 57

Comments comments comments…

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

9

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

10 of 57

Announcements

Case Study – A.1.1

Quick Topic Strengthening

[…Break…]

Course Project Consultation

​

from 18:00: Crit-Style Peer Review

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

10

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

11 of 57

General Thoughts

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

11

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

12 of 57

Let’s check out some examples…

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

12

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

13 of 57

Story-Telling

Many strongly compelling and engaging examples!

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

13

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

14 of 57

Reproducibility

Details on key add-ons must be specified.

e.g. what are the constraints you layered?

​

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

14

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

15 of 57

Reproducibility

Use your gallery post to expand on your details.

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

15

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

16 of 57

Continuity

If you expose an example at the beginning, try to follow-up on that same example.

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

16

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

17 of 57

Assessment

Clearly explain and show evaluation criteria.

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

17

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

18 of 57

Rigour & Consistency

  • What if you ask the same question again in a new AI chat instance?
  • What if you ask the same question with a different model? (e.g. Clause, Grok, etc)

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

18

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

19 of 57

Process Process Process

Singapore University of Technology & Design, Introduction to Design (2018)

Start drawing these flow diagrams!

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

19

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

20 of 57

Announcements

Case Study – A.1.1

Quick Topic Strengthening

[…Break…]

Course Project Consultation

​

from 18:00: Crit-Style Peer Review

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

20

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

21 of 57

Review

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

21

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

22 of 57

Example Tasks

Overall Goal

Mapping

Dataset

SUPERVISED

—Regression

—Classification

Pattern Recognition

f: X → y

X, y

UNSUPERVISED

—Clustering

—Dimensionality Reduction

Knowledge Disccovery

f: X → z

X

z: latent factors

REINFORCEMENT

... gaming, trading, layout editing, autonomous vehicle, robot control…

Sequential Decision Making

f: X(t) → X(t+1)

X(t), a(t), r(t), X(t+1)

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.02.25

22

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

23 of 57

Example Tasks

Overall Goal

Mapping

Dataset

SUPERVISED

—Regression

—Classification

Pattern Recognition

f: X → y

X, y

UNSUPERVISED

—Clustering

—Dimensionality Reduction

Knowledge Disccovery

f: X → z

X

z: latent factors

REINFORCEMENT

... gaming, trading, layout editing, autonomous vehicle, robot control…

Sequential Decision Making

f: X(t) → X(t+1)

X(t), a(t), r(t), X(t+1)

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.02.25

23

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

24 of 57

Machine Learning as Useful Mapping

IN

ℳ

OUT

x

y

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

24

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

25 of 57

Machine Learning as Useful Mapping

IN

ℳ

OUT

design of our structure

structural performance

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

25

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

26 of 57

Supervised Learning – Line Fit

ℳ

x

y

IN

x

y

OUT

IN

x

y

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

26

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

27 of 57

Unsupervised Learning - Clustering

x1,x2

z1,z2

[1,0] for points in Cluster 1

[0,1] for points in Cluster 2

algorithm determines what groups are

→

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

27

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

28 of 57

Data Points

1.0

1.0

 

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

28

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

29 of 57

Motivation: Modelling Diverse Non-Euclidean Designs

x1

x2

design distances

0.0

0.0

0.0

1.0

1.0

-1.0

0.0

-1.0

-1.0

-1.0

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

29

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

30 of 57

Motivation: Modelling Diverse Non-Euclidean Designs

x1

x2

distance

vector spaces allow us to represent data geometrically, making it possible to interpret similarity and difference in terms of distance.

0.0

0.0

0.0

1.0

1.0

-1.0

0.0

-1.0

-1.0

-1.0

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

30

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

31 of 57

Data Points

(0.0, 0.0)

0.0

0.0

x2

x1

 

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

31

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

32 of 57

Image as Data Matrices

https://ai.stanford.edu/~syyeung/cvweb/tutorial1.html

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

32

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

33 of 57

Machine Learning Life Cycle

  1. Gathering Data – Collect relevant raw data from various sources such as databases, sensors, APIs, or web scraping.
  2. Data Preparation – Clean and organize the data, removing duplicates and handling missing or inconsistent values.
  3. Data Wrangling – Transform and structure the data into a usable format suitable for analysis and model training.
  4. Analyse Data – Explore the data through visualization and statistical analysis to identify patterns and insights. Train
  5. Train Model – Use the processed data to train machine learning algorithms, allowing the model to learn from examples.
  6. Test Model – Evaluate the model’s performance using unseen data to measure its accuracy, precision, and generalization ability.
  7. Deployment – Integrate the trained model into production systems where it can make real-world predictions and be monitored for performance.
  • Machine Learning
  • Life Cycle

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

33

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

34 of 57

Data Preparation

x

y

x

y

or

?

x

y

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

34

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

35 of 57

Linear Regression

y=wx+b

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

35

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

36 of 57

Neural Networks Introduces Non-linearity

 

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

36

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

37 of 57

Representation Learning

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

37

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

38 of 57

On Distribution

Thought Experiment: �Let’s generate a typical Hong Konger…

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.03.04

38

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

39 of 57

On Distribution

What are key attributes that describe a person?

e.g. age, gender, income, height, weight….

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.03.04

39

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

40 of 57

On Distribution

Features live in data distributions with different shapes…

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.03.04

40

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

41 of 57

What do we mean when we are creating a generator?

Note: synthetic data generated by ChatGPT

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.03.04

41

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

42 of 57

Data Distribution

Conceptual Illustration of Data Generator

Note: synthetic data generated by ChatGPT

Where in the design space should our generator create its designs?

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.03.04

42

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

43 of 57

Valid Generated Data Point

Conceptual Illustration of Data Generator

Note: synthetic data generated by ChatGPT

✅ 👍 😃

This looks like a plausible generated data point.

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.03.04

43

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

44 of 57

Questionable Generated Data Point

Conceptual Illustration of Data Generator

Note: synthetic data generated by ChatGPT

❌ 👎 🤨

This is clearly outside of the distribution… not plausible.

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.03.04

44

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

45 of 57

Dimensionality Reduction by Manifold Learning : Concepts

H = W = 28

​

X ∈ ℝ H × W = 784

0

H

W

x1

x784

…

+

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.03.04

45

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

46 of 57

Dimensionality Reduction by Manifold Learning : Concepts

x1

x784

…

+

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.03.04

46

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

47 of 57

Variational Auto-Encoder

Dimensionality Reduction

ENC

z

 

xD

…

x1

x1

xD

…

+

+

+

+

+

+

+

+

+

+

+

+

+

+

DEC

 

Decoder

…

…

Encoder

z1

z2

+

+

+

+

+

+

+

+

+

+

+

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.03.04

47

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

48 of 57

Variational Auto-Encoder

Dimensionality Reduction

ENC

 

xD

…

x1

x1

xD

…

+

+

+

+

+

+

+

+

+

+

+

+

+

+

DEC

 

Decoder

…

…

Encoder

z1

z2

+

+

+

+

+

+

+

+

+

+

+

z

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.03.04

48

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

49 of 57

Modelling Diverse Non-Euclidean Designs

?

?

?

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

49

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

50 of 57

Text as Embedding Vectors

A long winded sentence…

…

ℳ

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

50

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

51 of 57

When can ML go wrong?

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

51

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

52 of 57

Real World Data

Real Relationship

Building model from limited realife data subset…

Training Data Scenario 1

Training Data Scenario 2

Training Data Scenario 3

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

52

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

53 of 57

Dtest

Dataset D

Dtrain

train on

Dtrain

test on

Dtest

ℳ

✂️

🫣

👁️

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

53

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

54 of 57

Validation via Dtest

Real-Life Data

Acquired Dataset

split into Dtrain and Dtest

​

​

Training via Dtrain

Model 1

Model 2

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

54

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

55 of 57

Attendance Check

Announcements

Case Study – A.1.1

Quick Topic Strengthening

[…Break…]

Course Project Consultation

​

from 18:00: Crit-Style Peer Review

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

55

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

56 of 57

Announcements

Case Study – A.1.1

Quick Topic Strengthening

[…Break…]

Course Project Consultation

​

from 18:00: Crit-Style Peer Review

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

56

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU

57 of 57

Thank You for Your Attention!

CCAI 9012 : GenAI Solutions to Global Challenges: Using AI Creatively and Responsibly

2026.04.01

57

Kam-Ming Mark TAM | Hongshan GUO | Waishan QIU