Introduction to AI

• Kecerdasan: kemampuan menyelesaikan masalah baru, berperilaku rasional, atau menyerupai manusia; melibatkan interaksi dunia nyata, reasoning/planning, serta learning/adaptation.
• AI menurut McCarthy: ilmu dan rekayasa membuat mesin cerdas; intelligence = bagian komputasional dari pencapaian tujuan di dunia.
• Empat tipe sistem AI: berpikir seperti manusia (cognitive science), bertindak seperti manusia (Turing test), berpikir rasional (logika), bertindak rasional (utility/decision theory).
• Sejarah singkat: McCulloch–Pitts (1943), Turing (1950), Dartmouth (1956), GPS & LISP, AI winter & expert systems (DENDRAL, MYCIN), revival neural nets (1986+), Bayesian networks, integrasi ML–reasoning–KR.
• Contoh sukses: Deep Blue, bukti Robbins conjecture, logistics Gulf War, autonomous spacecraft planning, Proverb crossword, DARPA driving challenge, face recognition kamera konsumen.
• Disiplin pendukung: filosofi, matematika, probabilitas/statistik, ekonomi, neuroscience; hardware otak jauh lebih paralel dibanding CPU.
• State of practice: speech/vision terbatas vocabulary masih kuat; scene understanding umum masih sulit; planning optimal sulit karena kombinatorial.
• Paradigma learning: supervised (label), unsupervised (struktur tersembunyi), reinforcement (reward kumulatif tanpa pasangan I/O eksplisit).

Slide memetakan empat definisi AI; diagram ini merangkum hubungan think vs act dan human vs rational.

flowchart TB
  subgraph human["Human-oriented"]
    T1["Think like humans"]
    A1["Act like humans"]
  end
  subgraph rational["Rational-oriented"]
    T2["Think rationally"]
    A2["Act rationally"]
  end
  T1 --> A1
  T2 --> A2
  human --- rational