AI · Theory

Expert Systems

Theory

48/215

Expert Systems

An expert system encodes a specialist’s rules: if symptoms then hypothesis. Knowledge base + inference engine + explanation module. MYCIN (old medical) is the history example. Brittle when the case is new. Today often mixed with ML.

Fresher — rules you can read, plus “why?” explanation. That explainability is the selling point vs a black-box net.

Expert Systems — output — recommend check measles / because fever+rash. Readable why.

Exam tip

Three parts of an expert system.

Example

# Tiny expert rule
rules = [("fever+rash", "check measles")]
facts = {"fever+rash"}
for cond, rec in rules:
    if cond in facts:
        print("recommend:", rec)
        print("because:", cond)

Expert Systems — output: recommend check measles / because fever+rash. Readable why.

Short notes

DefKB + inference + explanation.
RuleBrittle on novel cases.

AI · Theory

Expert Systems

← All stacks

Theory

48/215

Expert Systems

Fresher — rules you can read, plus “why?” explanation. That explainability is the selling point vs a black-box net.

Expert Systems — output — recommend check measles / because fever+rash. Readable why.

Exam tip

Three parts of an expert system.

Example

# Tiny expert rule
rules = [("fever+rash", "check measles")]
facts = {"fever+rash"}
for cond, rec in rules:
    if cond in facts:
        print("recommend:", rec)
        print("because:", cond)

Expert Systems — output: recommend check measles / because fever+rash. Readable why.

Short notes

DefKB + inference + explanation.
RuleBrittle on novel cases.

Expert Systems

Example

Short notes

Questions

Expert Systems

Example

Short notes

Questions