1. Goals of AI

The main goals of Artificial Intelligence are to replicate human intelligence, solve complex knowledge-based problems, and connect perception with action. AI also aims to develop systems that can learn, reason, plan, and perform tasks like humans.

Example: Self-driving cars that can learn driving patterns and make decisions on roads.

NOT FOR EXAMINATION:
AI tries to mimic human brain functions such as thinking, learning, and decision-making.
It is designed to reduce human effort and improve efficiency in real-world tasks.

Real-time Analogy: Like a smart assistant (Google Assistant) that understands your voice, thinks, and gives correct responses.


2. Simple vs Model-Based Agents

Simple reflex agents act only on current input using condition-action rules, whereas model-based agents use memory and an internal model to make better decisions based on past and present data.

Example:

  • Simple → Light turns ON when switch is pressed
  • Model-based → Smart thermostat adjusts temperature based on past usage
NOT FOR EXAMINATION:
Model-based agents are more intelligent because they understand the environment, while simple agents react blindly without memory.

Real-time Analogy: Simple → Automatic door opens | Model-based → Google Maps uses past + current traffic


3. Artificial Intelligence

Artificial Intelligence is a branch of computer science that creates systems capable of performing tasks like learning, reasoning, problem solving, and decision making similar to humans.

Example: Chatbots used in customer service

NOT FOR EXAMINATION:
AI is not just robots; it includes software systems like recommendation engines, chatbots, etc.

Real-time Analogy: ChatGPT answering like a human tutor


4. Problem Solving Steps

Steps include problem definition, data collection, model selection, solution generation, evaluation, and refinement.

Example: Predicting student marks using past data

NOT FOR EXAMINATION:
This structured approach helps AI find optimal solutions efficiently.

Real-time Analogy: Understand → Solve → Check → Improve


5. Knowledge Representation

Two types: logical representation and semantic networks.

Example:

  • Logical → “All birds can fly”
  • Semantic → Relationship diagrams
NOT FOR EXAMINATION:
Different techniques are used depending on complexity.

Analogy: Logical → “Humans are mortal” | Semantic → Family tree


6. Knowledge-Based Agents

They use a knowledge base (facts + rules) and reasoning mechanism to make decisions.

Example: Medical expert system

NOT FOR EXAMINATION:
They continuously update knowledge for better decisions.

Analogy: Doctor diagnosing disease


7. Types of Knowledge
  • Factual → Water boils at 100°C
  • Procedural → Steps to solve math
  • Heuristic → Shortcuts
NOT FOR EXAMINATION:
Heuristic knowledge is faster but not always perfect.

Analogy: Facts, Steps, Tricks


8. First Order Logic (FOL)

Includes objects, predicates, variables, functions, and quantifiers (∀, ∃).

Example: ∀x Human(x) → Mortal(x)

NOT FOR EXAMINATION:
FOL is more powerful than propositional logic.

9. Unification Applications

Used in theorem proving, NLP, expert systems.

Example: Chatbot matching queries

NOT FOR EXAMINATION:
Helps match patterns and variables.

10. Planning

A plan is a sequence of actions to achieve a goal.

Types: Total order, Partial order

NOT FOR EXAMINATION:
Planning helps decide next steps.

11. Unification Conditions

Requires matching predicates, same arguments, and consistent substitutions.

NOT FOR EXAMINATION:
Incorrect matching causes errors.

12. AI Planning Applications

Used in robotics and scheduling.

NOT FOR EXAMINATION:
Used in automation systems.

13. Partial Order Planning

Actions are flexible and not strictly ordered.

NOT FOR EXAMINATION:
Allows parallel execution.

14. Planning with Limited Resources

Handles constraints like time, money, resources.

NOT FOR EXAMINATION:
Efficient allocation is key.

15. Steps in Partial Order Planning

Goal → Action → Constraints → Conflict resolution → Refinement

NOT FOR EXAMINATION:
Avoids unnecessary restrictions.

16. Uses of Planning Draft

Helps organize tasks, reduce wastage, improve efficiency.

NOT FOR EXAMINATION:
Acts as blueprint.

17. Probability Theory

Used to handle uncertainty and predict outcomes.

Example: Rain probability 70%

NOT FOR EXAMINATION:
Used when exact answers are not possible.

18. Bayesian Reasoning

Updates probability using evidence (Prior + Posterior).

NOT FOR EXAMINATION:
Improves accuracy.

19. Bayes’ Theorem

Calculates updated probability using evidence.

NOT FOR EXAMINATION:
Widely used in ML.

20. Conditional Independence

Variables are independent given a third variable.

NOT FOR EXAMINATION:
Reduces complexity.