Artificial Intelligence

Dr. Nguyễn Văn Vinh

CS Department, Information Faculty-
UET

1
Introductor
• Dr. Nguyễn Văn Vinh
• PhD (2009) in JAIST (Japan Advanced Institute of Science and Technology), Japan
• Research focus: AI, Natural Language Processing, Large Language Models
• Lecturer, AI4Life lab leaders - UET– Hanoi VNU
• Senior Data Scientist (Leader of LLM applied domain group), AI Center, Fsoft
• Design Leader of Machine Learning course for FUNiX – Online University
(https://vnexpress.net/giao-duc/funix-ra-mat-chuong-trinh-machine-learning-
3986011.html)
• Principle Investigator of National KC Project: “Phát triển hệ thống dịch đa ngữ
giữa tiếng Việt và một số ngôn ngữ khác”, 2020-2022.
• Principle Investigator of VNU Hanoi Project: “Research and optimizing large
language model and applying for Vietnamese question-answering
system”, 2023-2025.
• Principle Investigator of GHTK-UET project: “Name Entity Recogintion” and
“Intents Detection”
• Teaching some ML, NLP Lectures in Samsung, BigData Vin, Viettel, Tp bank, …

2
Artificial Intelligence

• Instructor: Dr. Nguyen Van Vinh, CS Department, Information

Faculty.
Email: [email protected], [email protected]
Tel: 0912263062
• Course web page: https://portal.uet.vnu.edu.vn/courses/ choose
• AI course.
• Up to date information
• Lecture notes
• Relevant dates, links, etc.
• References:
• Lecture Slides
• https://stanford-cs221.github.io/autumn2019/#coursework
• Đinh Mạnh Tường: Trí Tuệ Nhân Tạo. NXB Khoa Học Kỹ Thuật 2022
• Từ Minh Phương: Giáo trình Nhập môn Trí Tuệ Nhân Tạo, 2016
• [AIMA] Artificial Intelligence: A Modern Approach, by Stuart
Russell and Peter Norvig. (3nd ed)
3
Artificial Intelligence: A Modern Approach, 2020
(http://aima.cs.berkeley.edu/ )

4
Your Helpful TAs

• Nguyen Minh Quy, Talent Bachelor, CS Department, Information

• Master student Nguyen Quang Phuc

5
Artificial Intelligence

• Course overview: foundations of symbolic intelligent systems.

Agents, search, problem solving, logic, representation, reasoning,
machine learning, natural language processing & computer
vision.

• Prerequisites: Programming principles, discrete mathematics for

computing, software design and software engineering concepts.
Good knowledge of C++, Java, Python. Python required for
programming assignments.

• Grading: 30% for (midterm + homeworks/assignments ) +10%

for attendence + 60% for final

6
Policy & Practical issues

• Encourage discussion but assignments must be your individual work

• Codes copied from books or other libraries but be explicitly
acknowledged
• Sharing or copying codes is strictly prohibited.

7
Generative AI Policy

• You may use generative AI tools such as CoPilot and

ChatGPT, as you would use a human collaborator.
“This means that you may NOT directly ask
generative AI tools for answers or copy
solutions. You're required to acknowledge
generative AI tools as collaborators and include
a paragraph describing how you used the tool”.
• The use of generative AI tools to substantially complete
an assignment or exam (e.g., by directly copying) is
prohibited and will result in honor code violations. We
will be checking students' assignments to enforce this
policy.

8
AI Market Prediction to 2030

9
ARTIFICIAL INTELLIGENCE
MACHINE LEARNING
DEEP LEARNING /Generative AI

10
AI course of Stanford 2017

National
Economics and
Foreign Trade
University has a
major in
Computer
Science!!!

11
Why study AI?

12
 Some achievements of
Artificial Intelligence

13
 Google Translate & Vietgle
Translate

14
 AlphaGo
• 3-2016, Alphago
defeated Lee Sedol,
the Korean go
grandmaster 4-1,
made a surprised
both professional
goers and AI
researchers.

• 2017, AlphaGo Zero,

which beat the
original AlphaGo
system 100—0.

15
 Image Net (better than Human)
 The best image classification system declined from
26% error in 2011 to 3.1% in 2016 equivalent and
lower than humans.

16
 Image Net (better than Human)

17
 Virtual Assistant

 Apple's Siri, Google

Now, Microsoft Cortana,
Amazon Echo, và other
assistants.
 Thanks to breakthrough
innovation of speech
recognition (ASR) in the
last 7 years.
 However, ASR is only a
phase, and another
phase is understanding
the text, this is a very
difficult problem.

18
 ChatGPT (AI hotest topic)
 ChatGPT (Chat Generative Pre-trained Transformer) is
a chatbot launched by OpenAI in November 2022
 Generative Pre-trained Transformer 3 (GPT-3) is
an autoregressive language model that uses deep learning to
produce human-like text. Given an initial text as prompt, it will
produce text that continues the prompt.
 The architecture is a standard transformer network (with a
few engineering tweaks) with the (back then) unprecedented
size of 2048-token-long context and 175 billion parameters
(requiring 800 GB of storage).
 ChatGPT can fluently answer all the questions that users ask,
any domains. Besides, ChatGPT can also write code, write
poetry, compose music, write letters (documents), design and
even fix errors in programming.
 ChatGPT has surpassed 10 million users just 40 days after
its official launch, breaking all previous records (100 million
just 2 months).
19
 The New Era of Pre-
training/LLMs

20
AI jobs in 2020

21
22
Companies

23
Companies (Vietnam)

• VinAI/Vin BigData Academy

• FPT.AI
• QAI (Fsoft)
• Viettel-AI
• Zalo.AI
• …

24
Governments

Vietnam? (Itrithuc.vn), AI4Life, AI4VN

25
 AI is having real-world impact

• Public imagination
• Economy
• Politics
 AI is having real-world impact

• Public imagination
• Economy
• Politics
 AI is having real-world impact

• Public imagination
• Economy
• Politics
AI Strategy 2021 of Vietnam

• On January 26, 2021, the Vietnamese Government

issued a national strategy on artificial intelligence
http://datafilesbk.chinhphu.vn/file-remote-
v2/DownloadServlet?filePath=vbpq/2021/01/127.signed.pdf

29
What is AI?

• “AI is the science and engineering of making intelligent machines

which can perform tasks that require intelligence when performed
by humans …”

30
What is AI?

Views of AI fall into four categories (bốn loại):

Lý trí vs. con người (human are not perfect)
Hành động vs. Suy nghĩ

Thinking humanly Thinking rationally

Acting humanly Acting rationally

31
 Alan Turing - Father of AI
Alan Turing (OBE, FRS)
• Born 23 June 1912, Maida Vale,
London, England
• Died 7 June 1954 (aged 41),
Wilmslow, Cheshire, England
• Fields: Mathematician, logician,
cryptanalyst, computer scientist
• Institutions:
• University of Manchester
• National Physical Laboratory
• Government Code and Cypher
School (Britain's codebreaking
centre) Alan Turing memorial
• University of Cambridge statue in Sackville Park,
Manchester

32
Alan Turing

• The history of AI begins with the following articles:

• Turing, A.M. (1950), Computing machinery and intelligence, Mind,
Vol. 59, pp. 433-460.

33
Acting Humanly: The Turing Test

• Alan Turing's 1950 article Computing Machinery and

Intelligence discussed conditions for considering a
machine to be intelligent

• “Can machines think?”  “Can machines behave

intelligently?”
• The Turing test (The Imitation Game): Operational definition of
intelligence.

34
Acting Humanly: The Turing Test

• Computer needs to possess: Natural language processing,

Knowledge representation, Automated reasoning, and Machine
learning

35
Branches of AI

36
Artificial Intelligence in diverse subject

37
38
39
40
41
 AI: a brief history (1)

42
 AI: a brief history (2)

43
 AI: a brief history (3)

44
 AI: a brief history (4)

45
 AI: a brief history (5)

46
 AI: a brief history (6)

47
 AI: a brief history (7)

48
 AI: a brief history (8)
 2000-present: Statistics
 Machine Learning
 Traditional models: Hidden Markov models (HMM), support
vector machines (SVM), Gaussian processes, graphical
models (Bayes networks, conditional random fields)
 Data mining
 2010- present: Deep Learing

49
Characteristics of AI tasks

• High societal impact (aect billions of people)

• Diverse (language, games, robotics)
• Complex (really hard)

51
Resources

52
Responsible AI & Ethic AI

• Understand, protect, and control your data, models, and processes

to build trusted solutions.
• Ethical AI is artificial intelligence that adheres to well-defined
ethical guidelines regarding fundamental values, including such
things as individual rights, privacy, non-discrimination (không phân
biệt đối xử), and non-manipulation (không thao túng).
• Ethical AI places fundamental importance on ethical considerations
in determining legitimate and illegitimate uses of AI.

53
Course Topics

Probability &
Search
Reasoning

Supervised Reinforcement
Learning Learning

Applications

Impact on Sciences, Technology, Society

Course Overview
• Problem Solving using search techniques:
• Blind search.
• Informed search
• Constraint satisfaction problem.
• Game Playing:
• Adversarial Search
• Knowledge Representation:
• Propositional logic
• Predicate logic
• Reasoning mechanisms
• Probabilistic reasoning
• Machine Learning:
• Machine Learning e.g. Naive Bayes, neural network, deep learning,
Reinforcement Learning
• Natural Language Processing & Computer Vision
• Reinforcement Learning
• Responsible AI & Ethic AI

55
Summary

• There’s no magic in AI. It’s all about math, computer science, and
smart ways to relate them to real problems
• AI is a fast-moving exciting area
• We can directly make the world a better place
• This course will teach you the foundations

56