You are on page 1of 49

ARTIFICIAL

INTELLIGENCE

Nguyễn Lê Hồng Hạnh


Mai Nguyễn Anh Vũ
Lương Tấn Khang
What is AI?

CONTENT A brief history of AI

Some Applications of AI

2
What is
Artificial Intelligence?

3
“ The science and engineering of
making intelligent machines,
especially intelligent computer
programs.
John McCarthy – The father of AI

4
“ … to be called intelligent, a
machine must produce responses
that are indistinguishable from
those of a human.
Alan Turing
The father of Computer Science

5
Artificial Intelligence is a way of making a
computer, a computer-controlled robot, or
a software think intelligently, in the similar
manner the intelligent humans think.
Robot Sophia (Source: Wiki) Vision in Avenger film
Autonomous car (Source: cnn.com)
Object Detection (Source: deepai.com)
The difference between
AI, ML and DL

AI ML DL
Artificial Intelligence: all things that make a
computer programme execute a job or
activity without the support of people.

A* algorithm (Source: geeksforgeeks.com)


Artificial Intelligence: all things that make a
computer programme execute a job or
activity without the support of people.

DeepBlue computer won the Master of Chess


Machine Learning: is to enable machines to
learn by themselves using the provided data
and make accurate predictions.

Deep Learning: is roughly inspired by the


information processing patterns found in the
human brain. Just like we use our brains to
identify patterns and classify various types
of information, deep learning algorithms can
be taught to accomplish the same tasks for
machines.
The main history of AI

1955 1986 1997 2016

Logic theorists, a first Deep Blue won the


AI programme is Master of Chess, Gary
invented. Add title Kasparov
Add title
Navlab, a first
Google DeepMind’s
autonomous car is
AlphaGo thắng nhà vô
invented by Carnegie
14
địch cờ vây Lee Sedol
Melon
The main history of Machine Learning

The historical development of ML


15 (source: http://www.erogol.com/brief-history-machine-learning/)
The main history of Deep Learning

The historical development of Deep Learning


16 (source: https://beamandrew.github.io/deeplearning/2017/02/23/deep_learning_101_part1.html)
Some current
applications of
Machine Learning

17
Types of Machine Learning

• Supervised learning
• Unspervised learning
• Semi-supervised learning

18
19
Reinforcement learning is an area of machine learning
concerned with how software agents should take
actions in an environment to maximize some notion of
cumulative reward.

20
Reinforcement learning is an area of machine learning
concerned with how software agents should take
actions in an environment to maximize some notion of
cumulative reward.

21
Some current
applications of
Deep Learning

22
Natural Language
Processing
A programme is capable of
comprehensive reading and
communicating in natural language
with people.

23
Text Classification
Text classification also known as text tagging or text
categorization is the process categorizing text into organized
groups.

24
Text Classification

25
Text Classification

26
Machine Translation

27
Machine Translation

28
Document summarization
Automatic summarization is the process of
shortening a text document with software, in order
to create a summary with the major points of the
original document.

Hình 8. Ví dụ nhận tóm tắt văn bản.


29
Document summarization

Link demo: https://resoomer.com/en/

30
Question Answering
Question Answering is a programme that
automatically answer questions posed by
humans in a natural language.

Question Answering has also been used to


develop dialog systems and chatbots
designed to simulate human conversation

Demo: http://35.165.153.16:1995/

31
Computer Vision
Computer vision is a field of study that
seeks to develop techniques to help
computers “see” and understand the
content of digital images/ videos.

32
Object detection

33
Object detection

34
Image Segmentation

35
Face Verification

36
Face Verification: One-shot Learning

Features Map ?

Features Map

Features Map
Compare

Features Map

37
Face Verification: One-shot Learning

38
Emotion Detection

Demo: https://aidemos.microsoft.com/face-
39 recognition
Landmark Recognition

40
Neural Style Transfer

41
Problem of Machine Learning and Deep Learning
A huge number of datasets

42
Solutions?
• Rent many people to annotate images.

• Data augmentation.

• Using GANs.

43
Generative Adversarial Networks (GANs)
GANs are a set of models that basically learn to create
synthetic data that is similar to input data it’s given,
including:

• Generator
• Discriminator

44
Generative Adversarial Networks (GANs)

Source (medium.com)
45
Image Resolution

46
GANs in Style transfer

Source: paper (https://arxiv.org/pdf/1706.03319.pdf)


47
GANs in Style transfer

Demo: https://paintschainer.preferred.tech/index_en.html

48
THANKS!
Any questions?

49

You might also like