COMP 2211 Exploring Artificial Intelligence

Introduction to Artificial Intelligence

Dr. Desmond Tsoi & Dr. Huiru Xiao
Department of Computer Science & Engineering
The Hong Kong University of Science and Technology, Hong Kong SAR, China
What is Artificial Intelligence?
There is not a widely accepted definition of artificial intelligence.
Alan Turing defines a favourable definition of Artificial Intelligence (AI):

AI is the science and engineering of making intelligent machines, especially intelligent

computer programs.

From our observations, AI borrows characteristics from human intelligence and applies
them as algorithms in a computer-friendly way.

Founding father of AI, Alan Turing (also the father of theoretical computer science)
desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 2 / 33
Why Study Artificial Intelligence (AI)?

Brighter career
You can work as a Machine Learning Engineer, Data Scientist, Business Intelligence Developer,
Research Engineer, etc., once you have studied AI.
AI is versatile (skillful)
AI is not just limited to the computer and space-related industries. Still, it also plays a significant
role in industries directly related to a common person, like healthcare, automobile, and even the
banking and finance sector.
Skill of the century
It is undeniably true that someday, AI will replace many jobs that humans do and create many
different job opportunities in related fields.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 3 / 33

Academic Disciplines to AI
AI is interdisciplinary, and its foundations are in various fields.
Philosophy and cognition science: Logic, methods of reasoning, foundation of learning,
language, rationality.
Mathematics: Important to write the logic and algorithm for machine learning. Good
knowledge of mathematics is a must skill to develop a model of AI.
Neuroscience: Provide information about how the human brain works and how neurons
respond to a particular event. This enables AI scientists to develop programming models
to work like the human brain.
Psychology: Study and find the process of thinking of humans and animals.
Computer Science: Write the codes for making the neural network for artificial
intelligence. Computer programmers should have very high programming skills and
knowledge of Mathematics and other disciplines used with AI.
Linguistics: Knowledge representation, grammar.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 4 / 33

Importance of AI
Gaming and entertainment: AI guides the actions of non-human player characters.
Speech recognition: AI converts speech to text.
Understanding natural language: AI carries out a dialogue with a computer using natural
language.
Computer vision: AI identifies and locates objects in digital images and videos.
Expert systems: AI emulates the decision-making ability of a human expert.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 5 / 33

Importance of AI (Continued)
Healthcare: AI assists diagnosis, provides telemedicine, assists surgery using robots,
monitors vital stats.
Air transport: AI delivers safer air transport while reducing its environmental impact.
Banking and Financial Institutions: AI automates services, manage risk.
Logistics: AI delivers goods using autonomous vehicles.
E-commerce: AI personalizes content for customers to boost sales. AI filters spam and
fake reviews.
Hiring: AI goes through thousands of CVs and ascertains if there’s a good fit.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 6 / 33

Confusing Terms

The terms Artificial Intelligence, Machine Learning, Neural Networks, Deep Learning tend
to be used interchangeably in conversation, leading to some confusion around the nuances
between them. But how are they related?
Machine learning (ML) is a sub-field of
artificial intelligence that provides systems
with the ability to automatically learn and
improve from experience without being
explicitly programmed.
In ML, there are different algorithms (e.g.,
neural networks) that help to solve problems.
Deep learning (DL) is a sub-field of machine
learning, which uses neural networks to
analyze different factors with a structure
similar to the human neural system.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 7 / 33

Advantages of AI
Reduction in human error: Computers do not make mistakes if they are programmed
properly.
Takes risks instead of humans: Overcome many risky limitations of humans by developing
an AI Robot which can do the risky things for us.
Available 7x24: Using AI, we can make machines work 7x24 without any breaks, and they
don’t even get bored, unlike humans.
Helps in repetitive work: Using AI, we can productively automate mundane tasks and can
remove “boring” tasks for humans.
Digital assistance: Use digital assistants to interact with uses which save the need for
human resources.
Faster decisions: Using AI alongside other technologies, we can make machines make
decisions faster than a human and carry out actions quicker.
New inventions: AI is powering many inventions in almost every domain which will help
humans solve the majority of complex problems.
desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 8 / 33
Disadvantages of AI
High costs of creation: As AI updates every day, the hardware and software need to get
updated with time to meet the latest requirements. Machines need repairing and
maintenance, which need plenty of costs.
Making humans lazy: AI is making humans lazy with its applications automating most of
the work.
Unemployment: AI is replacing most repetitive tasks and other works with robots.
Human interference is becoming less, which will cause a major problem in the
employment standards. Every organization is looking to replace the minimum qualified
individuals with AI robots that can do similar work with more efficiency.
No emotions: Machines can not develop a bond with humans, which is an essential
attribute for Team Management.
Lacking out of box thinking: Machines can perform only those tasks which they are
designed or programmed to do. Anything out of that tends to crash or give irrelevant
outputs, which could be a major backdrop.
desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 9 / 33
Prerequisites for AI

A stronghold on Mathematics, namely discrete math, calculus, statistics and probability.

A good amount of experience in programming languages like Python, or Java.
A stronghold in understanding and writing algorithms.
A strong background in data analytical skills.
The will to learn machine learning languages.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 10 / 33

Brief History of AI

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 11 / 33

Different Periods

1950 - 1956: The birth of Artificial intelligence

1956 - 1974: Symbolic AI
1974 - 1980: The first AI winter
1980 - 1987: Boom
1987 - 1993: The second AI winter
1993 - 2011: AI (Finally achieved some of its oldest goals, e.g., it began to be used
successfully throughout the technology industry. The success was due to increasing
computer power)
2011 - Present: Deep learning, big data and artificial general intelligence

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 12 / 33

1950 [The birth of AI]

English mathematician, computer scientist

Alan Turing wrote a landmark paper titled
“Computing Machinery and Intelligence” that
asked: “Can machines think?”.
He felt that such machines would eventually be
constructed.
But he also realized a bigger problem: “How
would we know if we have succeeded?”.

Prediction
Predicated that by 2000, a machine might have a 30% chance of fooling a layman for 5
minutes.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 13 / 33

The Turing Test
A test to empirically determine whether a computer has achieved intelligence.

In these tests, one of the humans functions as the questioner, while the second human
and the computer function as respondents.
Using a specified format and context, the questioner interrogates the respondents within a
specific subject area. After a preset length of time or number of questions, the questioner
is then asked to decide which respondent was human and which was a computer.
desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 14 / 33
1956 [The birth of AI]
John McCarthy coined the term “artificial intelligence” as the topic of the Dartmouth
Conference, the first conference devoted to the subject.

John McCarthy: True Father of AI

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 15 / 33
1959 [Symbolic AI]
Arthur Samuel coined the term “machine learning”, reporting on programming a
computer “so that it will learn to play a better game of checkers than can be played by
the person who wrote the program”.

Arthur Samuel: Pioneer in the field of computer gaming and AI

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 16 / 33

1965 [Symbolic AI]
The first natural language processing computer program, ELIZA, was created.

ELIZA, a computer program that emulates a Rogerian psychotherapist.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 17 / 33

1980s [Boom]
Edward Albert Feigenbaum developed the first expert system in artificial intelligence.

Edward Albert Feigenbaum: The father of expert systems

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 18 / 33

1990s [AI]
Cynthia Breazeal, a pioneer of social robotics and human–robot interaction, created the
robot Kismet at MIT in 1990s.
Kismet was designed to engage in social interaction with humans, marking a significant
step forward in the field of human-robot interaction.
Breazeal’s work has opened up new possibilities for how humans and robots can interact
and coexist.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 19 / 33

1997 [AI]

An IBM computer called IBM Deep Blue Beat the world chess champion, Garry Kasparov,
after a six-game match: two wins for IBM, one for the champion and three draws.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 20 / 33

2002 [AI]

iRobot launched Roomba, an autonomous vacuum cleaner that avoids obstacles.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 21 / 33

2009 [AI]

Google built the first self-driving car to handle urban conditions.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 22 / 33

2011 [AI]

IBM’s Watson supercomputer defeated humans in the final Jeopardy match.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 23 / 33

2011-2014 [DL, BD, Artificial General Intelligence]

Personal assistants like Siri, Google Now, Cortana use speech recognition to answer
questions and perform simple tasks.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 24 / 33

2014 [DL, BD, Artificial General Intelligence]

Ian Goodfellow comes up with Generative Adversarial Network (GAN).

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 25 / 33

2016 [DL, BD, Artificial General Intelligence]

AlphaGo beats professional Go player Lee Sedol 4-1.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 26 / 33

2018 [DL, BD, Artificial General Intelligence]

Most universities have courses in Artificial Intelligence, including the HKUST.

COMP 2211 Exploring Artificial Intelligence
(1st Offering: Spring 2022, 2nd Offering: Fall 2022,
3rd Offering: Spring 2023, 4th Offering: Fall 2023)
COMP 3211 Introduction to Artificial Intelligence
COMP 4211 Machine Learning
COMP 4221 Introduction to Natural Language Processing
COMP 4222 Machine Learning with Structured Data
COMP 4421 Image Processing
COMP 4471 Deep Learning in Computer Vision
COMP 4901V Large-Scale Deep Perception, Localization, and Planning for Autonomous
Vehicles

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 27 / 33

2022 [DL, BD, Artificial General Intelligence]
The release of ChatGPT (Chat Generative Pre-trained Transformer), a large language
model trained by OpenAI, demonstrated the enormous potential of AI for transforming
the way we communicate and interact with machines.

ChatGPT is an AI-powered language model developed by OpenAI, capable of generating

human-like text based on context and past conversations.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 28 / 33

2023 [DL, BD, Artificial General Intelligence]
The release of Bard.
Bard is an experimental conversational AI chat service developed by Google.
It operates in a similar manner to ChatGPT, but with a notable distinction: Bard retrieves
information directly from the web as a source of knowledge.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 29 / 33

Some Useful AI & Machine Learning Libraries

TensorFlow: It is a machine learning framework from Google. It is a fast, flexible, and

scalable open-source machine learning library for research and production. With
TensorFlow 2.0 and newer versions, more efficiency and convenience was brought to the
game.
Keras: It is the high-level application programming interface (API) built on top of
TensorFlow. It is incredibly user-friendly and easy to pick up. It is good for fast
experimentation.
PyTorch: It is TensorFlow’s direct competitor developed by Facebook, and is widely used
in research projects. It allows almost unlimited customization and is well adapted to
running tensor operations on GPUs.
Scikit-learn: It is another user-friendly framework that contains a great variety of useful
tools: classification, regressions and clustering models, as well as pre-processing,
dimensionality reduction and evaluation tools.

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 30 / 33

Comparisons

TensorFlow vs Pytorch: TensorFlow has a very difficult learning curve but has good
documentation and lots of examples available. PyTorch is slightly easier to use than
TensorFlow but its documentation is weak and skimpy.
Good choice for beginners: Scikit-learn and Keras.
desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 31 / 33
Pros and Cons of using Python for AI

Pros:
Easy to learn and read
A vast number of machine learning libraries and useful tools
Popularity and large community
Many useful machine-learning repositories
Easy to integrate with other programming languages
Cons:
Speed limitations
Not suitable for mobile and game development
Design limitations

desmond@ust.hk, huiruxiao@ust.hk COMP 2211 (Fall 2024) 32 / 33

 That’s all!
Any questions?