[go: up one dir, main page]
Scribd
Upload
11 views27 pages

Week-1-A - Introduction To AI

The document outlines a course on Artificial Intelligence (AI) taught by Dr. Abdul Baseer Buriro, focusing on both theoretical and practical foundations of AI and Machine Learning. It includes course objectives, assessment criteria, project requirements, and essential materials for study. Key topics cover AI concepts, rational agents, and the implications of AI in real-world scenarios.

Uploaded by

coham90828
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
11 views27 pages

Week-1-A - Introduction To AI

The document outlines a course on Artificial Intelligence (AI) taught by Dr. Abdul Baseer Buriro, focusing on both theoretical and practical foundations of AI and Machine Learning. It includes course objectives, assessment criteria, project requirements, and essential materials for study. Key topics cover AI concepts, rational agents, and the implications of AI in real-world scenarios.

Uploaded by

coham90828
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 27

ARTIFICIAL

INTELLIGEN
CE
ntroduction
INSTRUCTOR

 Dr. Abdul Baseer Buriro

 Research Domain: Applied Machine Learning and Deep Learning for EEG Signals, Signal
Processing
 R 204, AB-III, abdul.baseer@iba-suk.edu.pk
CONTENT

 Goal: To understand the theoretical and practical foundations and concepts of Artificial
Intelligence and Machine Learning
 CLOs:
 Understand key concepts and principles of AI
 Implement classical AI techniques

 Topics:
 AI, Agents and Environments
 Search algorithms/strategies
 Complex Environment, Adversarial Search, and Uncertainty
 Machine Learning Fundamentals
ASSESSMENTS

 Midterm – 30%
 Quizzes – 10%
 Project/Assignment – 10%
 Final – 50%
PROJECT REQUIREMENTS

 Inform me about scope of your project to agree upon

 2 – 3 members per group

 Choosing an appropriate topic is part of your project

 Submit a 1-page proposal ONE Week before the midterm exam date

 Submit report & code ONE Week before the final exam date

 Report must take explicit reference to content of at least ONE Week lectures

 Up to max 10 pages including figures tables references, no predefined format


PROJECT REPORT

Write a report describing


 The task and dataset used

 Any data pre-processing (if needed)

 The architecture of your agent/machine learning model

 Training objective, model selection, agent design space choices, etc.

 The performances of your agent/machine learning model

 How you applied the different techniques taught in class in your project

 The main challenges you faced


MATERIALS AND CREDITS

 Books:
 Russel, Norvig: Artificial Intelligence: A Modern
Approach https://aima.cs.berkeley.edu/
 Sebastian R. & Vahid M.: Python Machine
Learning
 Poole, Mackworth: Artificial Intelligence:
Foundations of Computational Agents.
https://artint.info/
 Tutorials:
 The Python Tutorial
https://docs.python.org/3/tutorial/
 IDE:
 Google Colab https://colab.research.google.com/
PREREQUISITES

 Basic Computer Science Skills


 Variables, loops, functions, classes

 Basic Mathematics Skills


 Linear algebra, derivative, probability

 If you haven’t written Python code before,


 https://docs.python.org/3/tutorial/

 Quick Math refresher (Chapters: 1–4)


 https://www.deeplearningbook.org/
THIS CHAPTER (WEEK) ‘LL COVER

 Introduction
 The Foundation of AI
 The History of AI
 The State of Art
 Risks and Benefits of AI
SO, WHAT IS AI?

So... what is AI??


10
WHAT IS ARTIFICIAL INTELLIGENCE?

 Researchers have persuaded several versions of AI. Broadly


 Systems that think like human (think humanly): Thought process
and reasoning
 Systems that act like human (act humanly): Behavior
 Systems that think rationally: Correct conclusions given correct
evidences
 Systems that act rationally: Doing things right
ARTIFICIAL INTELLIGENCE
THINK HUMANLY

 Three ways to determine, how human think?


 Introspection (trying to catch own thoughts as they go by)
 Psychological experiments (observing others in action)
 Brain imaging (observing the brain in action)
 Introspection + Psychological Experiments + Brain Imaging =
Cognitive Science
 By the help of cognitive science, we can build a machine that can
think like a human.
THINK HUMANLY
ACT HUMANLY
 Turing Test (1950)
 Operational definition of
intelligent behavior
 Can a human interrogator tell
whether (written) responses
to her (written) questions
come from a human or a
machine?
THINK RATIONALLY
Example  The ones on the right are measured
Alay is a man. against an ideal performance measure
All men are mortal.
Therefore, Alay is  Rational machine should be able to
mortal. develop a logic
Is Alay
mortal?
THINK RATIONALLY

 These laws of thoughts govern the operation of the mind, called


Logic
 Logic helps to build intelligent systems
 NOTE:
 Logic should be correct
 Solving problem with just few hundreds of facts can exhaust the
computational resources
 Therefore, there should be reasoning as which step should be taken first
ACT RATIONALLY
Rational Agent acts to achieve the best expected outcome
under uncertainty.

Sensors (various
cameras, Actuators (various
microphones, motors, pistons,
infrared range finder pumps etc.)
etc.)
ACT RATIONALLY

 Operate independently (autonomously)


 Perceive the environment
 Adapt to change
 Create and pursue the goal
 Therefore, intelligent machines need
 Good logic/inference
 Knowledge representation and reasoning
ACT RATIONALLY

Understa
Sense Think Act
nd
RATIONAL AGENT

 An agent is anything that can


be viewed as:
 Perceive its environment
 Act in that environment
 Hardware, software, or both
 Dog, thermostat, airplane,
 Does the right thing under given Perceiv
circumstances Think Act
e
RATIONAL AGENT – EXAMPLE
RATIONAL AGENT – EXAMPLE

 Rationality depends on PEAS:


 Performance  success criterion
 Environment
 Actuator(s)
 Sensor(s)
WHAT IS ARTIFICIAL INTELLIGENCE?
Thinking Humanly Thinking
Rationally

Acting Humanly Acting Rationally


BENEFICIAL MACHINES

 The standard model is not good for the long run


 it assumes we provided all objectives.
 For Example:
 In chess we can apply a standard model due to well-defined complete
objectives.
 In self-driving car model, we can not define all the objectives.

 Our preferences change to control the


steering, brake, acceleration, gear, etc to
drive safely.
 It is harmless to apply preference and
objectives in a simulation model but highly
risky in real-life scenarios.

25
BENEFICIAL MACHINES

 Machines can make noise to distract the chess


player, it is intelligence (increases chances to win)
but not the desired one.
 Machines are designed to achieve their objective but
to pursue our objectives.
 We need to transfer objectives perfectly
 If a machine does not know the objectives and
preferences then it has to act cautiously, ask
permission, learn more about our preferences
through observation, and defer to human control.
 Ultimately, we want agents that are provably
beneficial to humans
A STORY

 One day Joe Bear was hungry. He asked his friend


Irving Bird where some honey was. Irving told him
there was a beehive in the oak tree. Joe walked to
the oak tree. He ate the beehive.

[Shank, Tale-Spin System, 1984]

You might also like