Review

1
Review

• Ref Artificial Intelligence A Modern Approach
  ch. 1 s.d. 7, 9
• Rabu, 6 Mar 2002

2
Intro to AI

• Ada banyak definisi AI, dpt. dikelompokkan
  menjadi 4

3
Rational Agent

• Jenis-jenisnya a.l.
• Simple reflex agents.
• Agents that keep track of the world.
• Goal-based agents.
• Utility-based agents.

4
Properti Environment

• Environment properties
• Accessible vs. inaccessible.
• Deterministic vs. non deterministic.
• Episodic vs. nonepisodic.
• Static vs. dynamic.
• Discrete vs. continuous.

5
Problem Solving by Searching

• Secara kasar, simple problem-solving agent
• Percept ? Problem definition ? Search for
  solution ? Action.
• Problem types
• Single-state problem
• Multiple-state problem
• Contingency problem (contoh khusus antar 2 pemain
  interleaving problem).
• Exploration problem
• Problem definition
• Perlu abstraksi (menentukan state aksi yg.
  dipilih).
• Terdiri atas
• Initial state
• Operator
• Goal test
• Path cost

6
Searching

• Yg. perlu diperhatikan
• Completeness
• Time complexity
• Space complexity
• Optimallity
• Dapat digolongkan menjadi
• Uninformed search / blind search
• Informed search

7
Uninformed Search

• Uninformed search
• Breadth-first search (BFS).
• Uniform cost search.
• Depth-first search (DFS).
• Depth-limited search.
• Iterative deepening search (IDS).
• Bidirectional search.
• Perbandingan search di atas

8
Informed Search

• Macam informed search
• Best-first search
• Greedy search.
• A search.
• Memory bounded search
• Iterative Deepening A search (IDA).
• Simplified Memory-bounded A search (SMA).
• Iterative Improvement Algorithms
• Hill Climbing.
• Simulated Annealing.

9
Fungsi Heuristik

• Pemilihan fungsi heuristik
• Akurat, admissable.
• Efisien
• Membuat fungsi heuristik
• Melemahkan permasalahan
• Menggabungkan beberapa fungsi heuristik yg.
  admissable
• Statistik (tp. tdk. menjamin admissability).
• Memberikan bobot pada faktor-faktor yg.
  diperkirakan mempengaruhi tercapainya suatu goal
  (learning from observation)

10
Game Playing

• Game dpt. dibagi menjadi
• Algoritma search utk. game playing
• Minimax
• Minimax with Fixed Depth
• Alpha-Beta Pruning
• Expectiminimax

11
Knowledge and Reasoning

• Knowledge-based agent
• Lebih advance dari problem-solving agent
• Terdiri atas
• Harus dapat
• Merepresentasikan world, state, aksi, dsb.
• Menerima adanya informasi baru.
• Meng-update representasinya atas world.
• Menyimpulkan hidden property dr. world.
• Menyimpulkan aksi apa yg. akan dilakukan.

12
Logic

• Perlu representasi knowledge sedemikian sehingga
  dpt. diambil kesimpulan dari KB ? logic.
• Terdiri atas
• Knowledge representation
• Reasoning (proof theory)
• Knowledge representation didefinisikan
  berdasarkan
• Syntax cara penulisan sentence.
• Semantic arti dari sentence.
• Reasoning
• Inference procedure
• Soundness
• Completeness

13
Logic

• Logic yg. dibahas
• Propositional logic
• First Order Logic (First Order Predicate
  Calculus)
• Untuk memudahkan proses inference, dibentuk
  normal forms
• CNF
• DNF
• Horn form

14
Propositional Logic

• Syntax
• Constant logic (true, false).
• Propotition symbol.
• Gabungan 1/lebih propotition symbol dng. operator
  ? , ?, ?, ?, ? (berdasarkan presedence).
• Semantic
• Rules of inference
• Modus Ponens.
• And-Elimination.
• And-Introduction.
• Or-Introduction.
• Double-Negation Elimination.
• Unit Resolution.
• Resolution.

15
FOPC

• Syntax
• Constants, Predicates, Functions, Variables,
  Connectives, Equality, Quantifiers.
• Term constant variable function(term1,
  term2, ..., termn).
• Atomic sentence predicate(term1, ..., termn)
  term1 term2.
• Complex sentence Semantics
• Inference rule
• Inference rule utk. propositional logic
• Substitution
• Unification
• Generalized Modus Ponen
• Universal Elimination
• Existential Elimination
• Existential Introduction.

16
Proofing

• Macam proofing
• Forward chaining
• Backward chaining
• Resolution
• Resolution with Refutation
• Forward chaining
• Ubah problem stmt. jadi FOL
• Dari KB, gunakan inference rule yg. ada hingga
  terbukti hal yg. akan dibuktikan.
• Backward chaining
• Ubah problem stmt. jadi FOL
• Dari hal yg. mau dibuktikan (G), buktikan hal-hal
  yg. perlu dibuktikan agar G terbukti.

17
Proofing

• Resolution
• Ubah problem stmt. jadi FOL
• Ubah FOL jadi CNF clauses
• Lakukan resolusi dng. menggunakan KB yg. ada
• Resolution with Refutation
• Ubah problem stmt. jadi FOL
• Tambahkan negasi dari hal yg. akan dibutuhkan ke
  KB
• Ubah FOL jadi CNF clauses
• Lakukan resolusi dng. menggunakan KB yg. ada
• Resolusi yg. dilakukan dpt. secara forward /
  backward (forward chaining / backward chaining).

18
Mengubah FOLMenjadi Clause Form

• Eliminasi biimplikasi
• Eliminasi implikasi
• Reduce scope
• Standardize variables
• Skolemization
• Hilangkan universal quantifier
• Buat jadi CNF
• Buat jadi clauses
• Standardize variables lagi

19
Situation Calculus

• Representasi perubahan
• Effect axioms
• Frame axioms
• State-successor axioms
• Masalah
• Representational frame axioms
• Inferential Frame Problem
• Qualification Problem
• Ramification Problem