Week 1 - Intro to ai and search

Question 1

What does it mean for an AI system to “think like a human”?

Answer 1

The thought process and how it arrives at the outcome is like a human (may use intuition, emotions, or shortcuts).

Purpose: To replicate how humans reason and solve problems, including making mistakes and learning from experience.

Example: A human learning chess makes mistakes and improves over time through trial and error. Similarly, AlphaGo initially learned by simulating games against itself and adjusting after failures—this mimics human learning through experience.

If it were rational thinking, it would use optimization methods from the start to find the best moves without needing trial and error.

Question 2

Key things to remember about AI

Answer 2

AI is a type of intelligence (artificial, created by humans) used to solve problems.
Systems and behaviors can be artificially intelligent, but AI itself is not an object :
- The phrase ‘using an ai / an intelligence ‘ is consequently incorrect
It is also a scientific field focused on building intelligent systems

Question 3

What do we mean by INTELLIGENCE in a system

Answer 3

we cant tell if its a human or a machine

Question 4

What does it mean for an AI system to “act like a human”?

Back:

Answer 4

Definition: The system behaves indistinguishably from a human, regardless of how it reasons internally.
Purpose: To pass tests of human-like behavior.
Example: The Turing Test: In the original Turing Test, the human evaluator interacts with both the human and the AI, but they don’t know which is which. They ask questions to both and receive responses, after which they must decide which one is the human and which one is the machine.
Real-World Application: Chatbots that simulate human-like conversation in customer support.

Question 5

What is an agent

hint - 3 main features

Answer 5

. A unit that makes intelligent decisions when given the chance
. Uses knowledge it has at that point in time to make a decison
. Decision limited by range of actions possible at that point in time
-can only choose from the options its allowed to do even if it wants to act optimally

Question 6

What is an autonomous agent

Answer 6

agent that makes decisions without human input

ie self driving cars - they use sensors to navigate traffic without a human behind the wheel

Question 7

what is a rational agent

Answer 7

agent that always takes the best action at a given point

eg stockfish in chess

Question 9

AMAZING VIDEO FOR BEST FIRST SEARCH

Answer 9

https://youtu.be/Gbw1IsnY7KE?si=iEB2vl1S8n-upi2v

Question 10

difference between best first search and uniform cost search

Answer 10

Best-First Search (BFS): Expands the path with the fewest number of actions (edges).

Example: BFS will choose the path that takes the least number of road segments, regardless of the “travel cost” on each road.

In the diagram, BFS takes two steps (edges):
Sibiu → Fagaras → Bucharest (total cost: 310, actions: 2).

UCS:

Uniform Cost Search (UCS): Expands the path with the lowest total cost, where cost represents factors like distance, time, or fuel usage.

Example: UCS selects the path that minimizes the sum of travel costs (weights on the edges).

In the diagram, UCS chooses the path:
Sibiu → Rimnicu Vilcea → Pitesti → Bucharest (total cost: 278, actions: 3).