Practice Questions Collection

To pass the Total Turing Test, a computer system must demonstrate:

1. Natural Language Processing (NLP) - to communicate in natural human languages
2. Knowledge Representation - to store and retrieve general knowledge
3. Automated Reasoning - to use stored information to answer questions and draw new conclusions
4. Machine Learning - to adapt to new circumstances and detect patterns
5. Computer Vision - to perceive objects visually
6. Robotics (Motor Control) - to manipulate objects and move

Feature	Goal-Based Agent	Utility-Based Agent
Decision Basis	Achieves defined goal(s)	Maximizes a utility function
Evaluation	Binary (Goal achieved or not)	Graded (How good is the state)
Flexibility	Lower	Higher
Example	Path planning in maze	Route optimization with fuel cost

Solution Approach:

1. Identify start and goal nodes
2. Calculate f(n) = g(n) + h(n) for each node
3. Apply A* search algorithm
4. Track the path with lowest cumulative cost

Alpha-Beta Pruning: An optimization of minimax that eliminates unnecessary branches in game trees.

• Alpha (α): Best value MAX can guarantee
• Beta (β): Best value MIN can guarantee

Prune when α ≥ β. Example:

         A (MAX)
       /   |    \
     3    12     8 → Prune anything > 8 under MIN
                    

A Learning Agent improves its performance based on feedback from experience.

Components:

• Performance Element: Chooses external actions
• Learning Element: Makes improvements
• Critic: Evaluates performance
• Problem Generator: Suggests exploratory actions

Component	Description
Performance	Safety, speed, passenger comfort
Environment	Roads, traffic, weather conditions
Actuators	Steering, brakes, accelerator
Sensors	Cameras, LIDAR, GPS

Constraint Satisfaction Problem (CSP) has:

• Variables: Q1 to Q8
• Domains: 1 to 8 (columns)
• Constraints:
  • No two queens in same row
  • No two queens in same column
  • No two queens in same diagonal

Feature	Propositional Logic	Predicate Logic
Expressiveness	Low	High
Uses	Facts	Facts + Relationships
Example	P ∨ Q	Loves(John, Mary)

Planning Domain Definition Language (PDDL)

Spare Tire:

• Objects: spare, flat, axle
• Initial: flat on axle, spare in trunk
• Goal: spare on axle
• Actions: Remove(flat), Remove(spare), PutOn(spare)

Agent	Performance	Environment	Actuators	Sensors
Self-driving Car	Safety, speed	Roads	Wheels, brakes	Cameras, GPS
Vacuum Cleaner	Cleanliness	House floor	Wheels, vacuum	Dust sensor

Strategy	f(n)	Optimal	Informed
UCS	g(n)	Yes	No
Greedy BFS	h(n)	No	Yes
A*	g(n) + h(n)	Yes	Yes

• State space: All rover positions on Mars
• Initial state: Starting point of rover
• Actions: Move North, South, etc.
• Transition model: New position = old + action
• Path cost: Energy or time
• Goal test: At sampling location

Steps:

1. Initial Population
2. Selection (based on fitness)
3. Crossover
4. Mutation
5. New generation

Example: Maximize f(x) = (a+b)-(c+d)+(e+f)-(g+h)

Chromosomes: 87126601, 65413532...

Crossover at middle, mutation flips bit

• Percepts: Breeze, Stench
• Rules: If breeze, then pit nearby
• Represent in propositional logic:
  • B[1,1] → P[1,2] ∨ P[2,1]

Use resolution or model checking to infer safe paths.

Feature	Omniscient	Rational
Knows all	Yes	No
Chooses	Perfect decision	Best given knowledge

Applications:

1. ChatGPT (NLP)
2. Tesla Autopilot (Robotics)
3. Google Search (Inference)
4. Siri / Alexa (Voice Assistants)