Download Intro

Artificial Intelligence
Dr. Bashar Tahayna
COMP338
What is Intelligence?
• “Intelligence: The ability to learn and solve problems”
Webster’s Dictionary
What is AI ?
• Artificial intelligence (AI) is a branch of computer science that deals with
the creation of intelligent agents, which are systems that can reason,
learn, and act autonomously.
• AI research has been highly successful in developing effective techniques
for solving a wide range of problems, from game playing to medical
diagnosis.
• In this course, we will explore the foundations of AI, including search,
reasoning, planning, learning, and language processing. We will also
discuss the ethical and societal implications of AI.
The foundation of AI
Types of Artificial Intelligence
Artificial Intelligence can be divided based on:
• Capabilities
• Functionalities.
Capabilities-based AI Type
• Narrow AI (Weak AI):
• Focused on specific tasks, excels in performing a single function
(e.g., playing chess, recognizing faces, composing music). Limited
to the data and context provided.
• General AI (Strong AI):
• Hypothetical future AI with human-level intelligence or
surpassing it, capable of understanding and learning any
intellectual task that a human can. Currently not achieved.
• Superintelligence
Superintelligence):
(Artificial
• Hypothetical future AI surpassing human intelligence in all
aspects, potentially posing existential risks. A highly debated and
speculative concept.
The AI Umbrella
A MELDING OF
MINDS
ARTIFICIAL
INTELLIGENCE
MACHINE LEARNING
DEEP LEARNING
AI Applications
• Healthcare:
• Medical diagnosis, drug discovery, personalized medicine, robotic surgery.
• Finance:
• Fraud detection, algorithmic trading, risk assessment, personalized financial advice.
• Manufacturing:
• Predictive maintenance, automated quality control, logistics optimization.
• Transportation:
• Self-driving cars, traffic management, route optimization.
• Retail:
• Personalized recommendations, demand forecasting, customer service automation.
• Entertainment:
• Movie recommendations, music generation, video game AI.
• Environment:
• Monitoring climate change, optimizing energy usage, predicting natural disasters.
AI Applications- Examples
• AlphaFold:
• AI program that predicts protein structures, aiding drug discovery.
• DeepMind's AlphaZero:
• AI that mastered chess, Go, and Shogi, demonstrating learning across domains.
• Amazon Go stores:
• Cashierless shopping experience enabled by computer vision and sensor fusion.
• Tesla Autopilot:
• Driver-assistance system using vision and radar for semi-autonomous driving.
• Netflix recommendations:
• Personalized movie suggestions based on user preferences and viewing history.
Challenges and Opportunities of AI
• Challenges:
• Bias and Fairness:
• Ensure AI systems are unbiased and do not discriminate against specific groups.
• Transparency and Explainability:
• Understand how AI systems make decisions and ensure they are interpretable.
• Privacy and Security:
• Protect user data and privacy while ensuring the security of AI systems.
• Job displacement:
• Address potential job losses due to automation and prepare workforce for new roles.
• Ethical considerations:
• Ensure responsible development and use of AI aligned with ethical principles.
Challenges and Opportunities of AI
• Opportunities:
• Enhanced decision-making:
• Leverage AI for data-driven insights and improved decision-making across various fields.
• Automation and efficiency:
• Automate repetitive tasks and optimize processes for increased efficiency.
• Innovation and creativity:
• Utilize AI for new discoveries, inventions, and creative endeavors.
• Personalized experiences:
• Offer customized experiences and services tailored to individual needs.
• Tackling global challenges:
• Apply AI to address global issues like climate change, poverty, and disease.
AI Capabilities & the Underlying Ideas
• The Power of Combining Concepts
• AI capabilities are not built on a single idea, but rather a powerful
combination of several key concepts
The Power of Combining Concepts
Search and Problem-Solving:
• AI excels at searching through vast amounts of data to find the best
solution to a given problem.
• This involves algorithms like heuristic search, optimization
techniques, and game-playing strategies.
• Example:
• AlphaGo, an AI program, defeated human champions in the complex game of Go by
systematically searching for optimal moves.
The Power of Combining Concepts
Knowledge Representation and Reasoning:
• AI systems can represent and reason about the world using various
techniques, including logic, knowledge graphs, and probabilistic
models.
• This allows them to understand relationships between concepts and
draw inferences.
• Example:
• Medical diagnosis systems use knowledge graphs to link symptoms, diseases, and
treatments, aiding doctors in accurate diagnoses.
The Power of Combining Concepts
Handling Uncertainty and Incomplete Information:
• Real-world data is often messy and incomplete. AI systems can handle
this uncertainty using probabilistic inference and statistical methods
to make informed decisions even with limited information.
• Example:
• Self-driving cars use probabilistic models to assess their surroundings, navigate
uncertainties, and make safe driving decisions.
The Power of Combining Concepts
Optimization and Learning:
• AI systems can learn and improve over time by optimizing their
performance based on feedback or data.
• This involves techniques like machine learning, deep learning, and
reinforcement learning.
• Example:
• Recommendation systems learn your preferences and suggest personalized products
based on your past behavior and interactions.
The Power of Combining Concepts
Inspiration from Human Intelligence:
• Neural networks, inspired by the human brain's structure and
function, allow AI systems to learn complex patterns and relationships
from data.
• This enables tasks like image recognition, natural language
processing, and speech generation.
• Example:
• Deepfake technology uses neural networks to manipulate images and videos, creating
realistic but synthetic human faces.
The Power of Combining Concepts
Natural Language Processing (NLP):
• AI can understand and generate human language, enabling
applications like chatbots, machine translation, and sentiment
analysis.
• This involves techniques like tokenization, word embeddings, and
language models.
• Example:
• Virtual assistants like Siri and Alexa use NLP to understand your voice
commands and respond accordingly.
SEARCH
Initial State
ith State
Sequence of actions

Final State
Sequence of actions

SEARCH
Start
.
Correct sequence of actions

to achieve a goal
.
End
S
E
A
R
C
H
Search Problems - Terminology
• Agent:
• Entity that perceives its environment and acts upon that environment.
• State:
• A configuration of the agent and its environment
• Initial State:
• The state in which the agent begins
• Actions:
• Choices (or functions) that can be made in a state
Search Problems - Terminology
ACTIONS(s) returns the set of actions than can be executed in the
state s
Move down
Move right
Move left
Move up
Search Problems - Terminology
•So what is the relation between an action and a
state ?
Transition model
Search Problems - Terminology
Transition model
• A description of what state results from performing any applicable action in any state.
RESULT(s, a) returns the state resulting from performing action a in state s.
Transition model
RESULT(
,
)=
Search Problems - Terminology
State space
• The set of all states reachable from initial state by any sequence of actions
Search Problems - Terminology
Goal test
• A way to determine of a given state is a goal state.
Path cost
• Numerical cost associated with a given path.
Varying cost
4
2
2
5
1
3
4
2
6
3
6
1
2
2
Similar cost
2
2
2
2
2
2
2
2
2
2
2
2
2
2
Search Problems
• Initial states
• Actions
• Transition model
• Goal test
• Path cost function
Search Problems - Terminology
Solution
• A sequence of actions that leads from the initial state to a goal state.
Optimal Solution
• A solution that has the lowest path cost among all solutions.
Search Problems - Terminology
Node
• A data structure that keeps track of a state, a parent, an action, and a path
code. It represents a particular state in the search process.
This refers to the current configuration of the problem
being solved. It could be anything from the position of a chess
piece to the state of a Rubik's Cube.
Parent: This points to the node that generated the current
node. This helps trace the path taken to reach the current state.
Action: This specifies the action that was taken to transition from
the parent state to the current state.
Path Cost: This is the cumulative cost of the actions taken to reach
the current state from the initial state.
• State:
•
•
•
Approach
• Start with a frontier that contains the initial state.
• Repeat:
•
•
•
•
If a frontier is empty, then no solution.
Remove a node from a frontier
If a node contains goal state, return the solution.
Expand node, add resulting nodes to the frontier.
Example
Find a path from A to E
A
Frontier
B
C
D
• Start with a frontier that contains the initial state.
• Repeat:
• If a frontier is empty, then no solution.
• Remove a node from a frontier
• If a node contains goal state, return the solution.
• Expand node, add resulting nodes to the frontier.
E
F
Example
Find a path from A to E
A
Frontier
B
C
D
• Start with a frontier that contains the initial state.
• Repeat:
• If a frontier is empty, then no solution.
• Remove a node from a frontier
• If a node contains goal state, return the solution.
• Expand node, add resulting nodes to the frontier.
E
F
Example
Find a path from A to E
A
Frontier
B
C
D
• Start with a frontier that contains the initial state.
• Repeat:
• If a frontier is empty, then no solution.
• Remove a node from a frontier
• If a node contains goal state, return the solution.
• Expand node, add resulting nodes to the frontier.
E
F
Example
Find a path from A to E
A
Frontier
B
C
D
• Start with a frontier that contains the initial state.
• Repeat:
• If a frontier is empty, then no solution.
• Remove a node from a frontier
• If a node contains goal state, return the solution.
• Expand node, add resulting nodes to the frontier.
E
F
Example
Find a path from A to E
A
Frontier
B
D
C
• Start with a frontier that contains the initial state.
• Repeat:
• If a frontier is empty, then no solution.
• Remove a node from a frontier
• If a node contains goal state, return the solution.
• Expand node, add resulting nodes to the frontier.
E
F
Example
Find a path from A to E
A
Frontier
B
D
D
C
• Start with a frontier that contains the initial state.
• Repeat:
• If a frontier is empty, then no solution.
• Remove a node from a frontier
• If a node contains goal state, return the solution.
• Expand node, add resulting nodes to the frontier.
E
F
Example
Find a path from A to E
A
Frontier
B
D
D
C
• Start with a frontier that contains the initial state.
• Repeat:
• If a frontier is empty, then no solution.
• Remove a node from a frontier
• If a node contains goal state, return the solution.
• Expand node, add resulting nodes to the frontier.
E
F
Example
Find a path from A to E
A
Frontier
B
D
D
C
• Start with a frontier that contains the initial state.
• Repeat:
• If a frontier is empty, then no solution.
• Remove a node from a frontier
• If a node contains goal state, return the solution.
• Expand node, add resulting nodes to the frontier.
E
F