AI · Theory

Subsets of AI

Theory

47/215

Subsets of AI

AI ⊃ ML ⊃ DL is the usual picture. AI also includes search, logic, planning, robotics without any neural net. ML learns from data. DL uses deep nets. NLP, vision, RL sit across these. Generative models are a current DL/NLP slice.

Trap — “AI means neural networks only”. Search and expert systems are still AI.

Output — idea / do / check for Subsets of AI. Fill those three in the viva.

Diagram

AI
      └── ML
           └── DL
  (+ search, logic, robotics)

Exam tip

Draw the three nested circles + one outsider (search).

Example

# Subsets of AI
print("idea :", "Subsets of AI")
print("do   : one tiny example on paper")
print("check: one failure case")

Output: idea / do / check for Subsets of AI. Fill those three in the viva.

Short notes

DefAI ⊃ ML ⊃ DL
RuleSearch/logic also AI

AI · Theory

Subsets of AI

← All stacks

Theory

47/215

Subsets of AI

Trap — “AI means neural networks only”. Search and expert systems are still AI.

Output — idea / do / check for Subsets of AI. Fill those three in the viva.

Diagram

AI
      └── ML
           └── DL
  (+ search, logic, robotics)

Exam tip

Draw the three nested circles + one outsider (search).

Example

# Subsets of AI
print("idea :", "Subsets of AI")
print("do   : one tiny example on paper")
print("check: one failure case")

Output: idea / do / check for Subsets of AI. Fill those three in the viva.

Short notes

DefAI ⊃ ML ⊃ DL
RuleSearch/logic also AI

Subsets of AI

Example

Short notes

Questions

Subsets of AI

Example

Short notes

Questions