Please enable JavaScript.
Coggle requires JavaScript to display documents.
Knowledge-Based Systems and Environments - Coggle Diagram
Knowledge-Based Systems and Environments
Logical Agents
Pathfinding agents are limited.
Need Knowledge Base for expert systems.
Humans reason using internal knowledge.
Knowledge-Based Systems (KBS)
Captures human expert knowledge.
Supports AI decision-making.
Architecture
Knowledge Base: Facts and axioms.
Knowledge Representation Language (KRL)
First-Order Logic
Semantic Networks
Ontologies
Propositional Logic
TELL / ASK Operations:
Add information.
E.g., TELL the system "Socrates is a man.“
Query knowledge.
E.g., ASK "Is Socrates mortal?"
Inference
Deduce new facts from existing facts.
Example
All humans are mortal. It is raining
Wumpus World (Example of KBS Agent)
Problem Statement
A cave of 16 rooms (4×4 grid).
Dangers: Wumpus (monster) and bottomless pits.
Goal: Collect gold and escape safely.
Environment Clues
Breeze → Pit nearby
Glitter → Gold nearby
Stench → Wumpus nearby
Scream → Wumpus killed
P.E.A.S.
Environment
16 rooms cave, adjacent stench, breeze, glitter
Actuators
Move Forward, Turn Left/Right, Grab, Shoot Arrow, Climb
Performance measure
+1000 (Gold + Exit), -1000 (Death), -1 (Move), -10 (Arrow)
Smell, Feel (Breeze/Bump), See (Glitter), Hear (Scream)
Environment Properties
Deterministic: Actions have predictable outcomes.
Sequential: Series of actions impact future decisions.
Single-Agent: Only one agent (Wumpus considered environment).
Static: Wumpus and pits do not move.
Partially Observable: Limited perception.
Discrete: Limited, countable rooms and actions.
Knowledge Base & Propositional Rules
Example Rules
Stench(i,j) ⇔ Wumpus adjacent
Breeze(i,j) ⇔ Pit adjacent
OK(i,j) ⇔ (¬Wumpus ∧ ¬Pit)
Simple KB Facts
Stench at (1,2) ⇒ Wumpus nearby.
Glitter at (2,2) ⇒ Gold present.
No breeze at (1,1) ⇒ Adjacent rooms safe.
Agent Plan Example
Start at (1,1) → Move to (1,2) → Move to (2,1) → Move to (2,2) → Grab Gold → Return to (1,1).
