o Review Sebelum UTS

o Searching Solution : Informed Search => Case Game Tree (8 puzzle, Congklak)
o Pengantar Deep Learning
o Pengantar Pengenalan Pola
o Pengantar NLP
o Tren Penelitian Kecerdasan Artifisial
Peserta mampu menjelaskan konsep dasar AI dengan benar dan mengidentifikasi pemanfaatan
teknologi AI

o Definisi Artificial Intelligence

o Perancangan Intelligent Agent
o Jenis Teknologi Artificial Intelligence
o Perkembangan Teknologi Artificial Intelligence
o Penerapan Teknologi Artificial Intelligence
o Tantangan dalam Pengembangan Teknologi Artificial Intelligence
 ARTIFICIAL INTELLIGENCE, IS IT A HYPE?

Image Processing

Text Processing
Facebook Face Recognition
Speech Processing

Speech Recognition
AI Di Berbagai Negara
Arahan Teknologi Artificial Intelligence di Indonesia

Strategi Nasional Kecerdasan Artifisial

 McKinsey: Sept 2019
AI as Disruptive Technology Automation Effect in
Indonesia:

● Jobs replaced by Artificial Intelligence

technology are the repetitive one and can be
easily predicted. Most of AI Technology aims to
help human work
● Industry in Indonesia has started to use AI
technology to reduce cost, increase revenue,
giving added value to product
 Artificial Intelligence Agent

Artificial Intelligence

Cloud Computing
Agent

Big Data
Percept Action

IoT
Environ-
ment
 Internet of Things
Definisi:

Peralatan/gadget sehari-hari atau objek yang

terhubung dengan Internet dan memiliki
kemampuan untuk mengumpulkan dan
mengirimkan data

Contoh IoT pada keseharian/rumah dalam bentuk

smart consumer goods: smart bulb, smart light
switch, smart doorlock, smart watch, smartphone,
fitness tracker, smart tv, smart thermostat, smart
toilet, smart bike lock dll.
 Cloud & Edge Computing

o Menyimpan dan memproses data di komputer (datacenter)

orang lain melalui jaringan Internet
o Pemrosesan data di peralatan seperti smartphone (yang saat
ini semakin powerful)
o Sudah banyak dipakai di kehidupan sehari-hari
o Saat mengakses e-mail berbasis Web (misal Gmail)
o Saat upload foto/video di Facebook/Youtube/Google Drive
o Saat menggunakan Office 365, Google Doc
o Saat order gojek atau booking hotel lewat apps di
smartphone
o Untuk bisnis, platform untuk consumer services, inventory
Sumber gambar: https://www.simplilearn.com/tutorials/cloud-computing-
management, recruiting & HR, design, retail dan shipping tutorial/cloud-computing-architecture
(oleh cloud provider kebanyakan sudah disediakan sebagai
“software-as-a-service”)
 Big Data

Data Mining Big Data

Statistics and
Probability Data Warehouse AI - Machine Learning

1900 1980 1990 2000 Today

 Big Data
• High-volume, high-velocity, and/or high-variety information assets
• Require new forms of processing: capture, curation, storage, search, sharing, transfer, analysis, visualization
• To enable:
• enhanced decision making
• insight discovery
• and process optimization

• Karakteristik Big Data

• Velocity: kecepatan data yang dihasilkan (per detik, per menit, per jam, per hari, dst).
• Volume: jumlah data yang diakumulasikan (terabyte, petabyte, exabyte, zettabye, yottabyte dst)
• Variety: jenis/ragam data yang bermacam-macam: terstruktur, semi-terstruktur, tidak terstruktur
(teks, suara, gambar, video dll)
• Veracity: kesesuaian dengan fakta dan akurasi (khususnya dari data tidak terstruktur)
• Value: kemampuan untuk mengubah data ke value (profit, manfaat medis & social, customer
satisfication)
 Ekosistem Big Data
Acquisition
Data Sources Consumption
and Storage Analytics

Batch
ERP,
CRM, Statistik Data
OLAP Pola
(korelasi, asosiasi)
Streaming Reasoning
Machine Learning
NLP
Distributed Speech
Storage Vision
 Analytics on Data
01
Descriptive:
Menjelaskan keadaan bisnis saat ini melalui data historis.

02 Diagnostic:
Menjelaskan mengapa suatu masalah terjadi dengan melihat data historis.

Predictive:
03 Memproyeksikan atau memprediksi hasil masa depan
berdasarkan data historis.

Prescriptive:
04 Menggunakan hasil analitik prediktif dan pengetahuan lain
dengan menyarankan upaya terbaik di masa depan.
 Definisi Artificial Intelligence
Thinking Humanly Thinking Rationally

“The exciting new effort to make computers think.. “The study of mental faculties through the use of
Machines with minds, in the full and literal sense” computational models.” (Charniak and McDermott,
(Haugeland, 1985) 1985)
“[The automation of] activities that we associate with “The study of the computations that make it possible
human thinking, activities such as decision making, to perceive, reason, and act.” (Winston, 1992)
problem solving, learning …” (Bellman, 1978)
Acting Humanly Acting Rationally

“The art of creating machines that perform functions “Computational Intelligence is the study of the design
that require intelligence when performed by people.” of intelligent agents.” (poole et al., 1998)
(Kurzweil, 1990) “AI .. Is concerned with intelligent behavior in
“The study of how to make computers do things at artifacts.” (Nilsson, 1998)
which, at the moment, people are better.” (Rich and
Knight, 1991)
Definisi Artificial Intelligence (2)
 Intelligent Agent

Merancang PEAS:
Agent - Performance/Pengukuran performansi:
aman, cepat, tidak melanggar aturan lalu
lintas, kenyamanan penumpang

Percept Action - Environment/Lingkungan: jalan, rambu-

rambu lalu lintas, kendaraan lain,
penumpang
Environ- - Actuator: kemudi, gas, rem, klakson
ment - Sensor: kamera, sonar, speedometer, GPS
 Environment Type
Fully observable (vs. partially An agent's sensors give it access to the complete state of the
observable) environment at each point in time.
Deterministic(vs. stochastic) The next state of the environment is completely determined
by the current state and the action executed by the agent. (If
the environment is deterministic except for the actions of
other agents, then the environment is strategic)
Episodic (vs. sequential) The agent's experience is divided into atomic "episodes"
(each episode consists of the agent perceiving and then
performing a single action), and the choice of action in each
episode depends only on the episode itself
 Environment Type (2)

Static (vs. dynamic) The environment is unchanged while an agent is deliberating.

Discrete(vs. continuous) A limited number of distinct, clearly defined percepts and
actions.
Single agent(vs. multiagent) An agent operating by itself in an environment
Known(vs Unknown) This distinction refers not to the environment itself but to the
agent’s (or designer’s) state of knowledge about the “laws of
physics” of the environment
 Example on Environment Type
Chess with a clock Chess without a clock Taxi driving

Fully Observable Yes Yes No

Deterministic Deterministic Deterministic No

Episodic No No No

Static Semi Yes No

Discrete Yes Yes No

Single agent No No No

• The environment type largely determines the agent design

• The real world is (of course) partially observable, stochastic, sequential, dynamic, continuous, multi-agent
 Pembagian Jenis Teknologi Artificial Intelligence (1)
o Berdasar task scope:
o Artificial Narrow Intelligence (ANI) – weak AI: teknologi AI yang ditujukan untuk melakukan satu task khusus
o Contoh: chatbot pemesanan tiket pesawat, klasifikasi jenis penyakit kulit
o Artificial General Intelligence (AGI) – strong AI: teknologi AI yang bisa menangani semua task yang dilakukan
manusia

Sumber gambar: analyticsindiamag.com

 Pembagian Jenis Teknologi Artificial Intelligence (2)
o Berdasar domain persoalan:
o Teknologi AI juga dapat dibagi menjadi cabang-cabang domain persoalan yang diselesaikan, contohnya: Natural
language processing (input/output berupa Bahasa), Speech Processing (input/output berupa sinyal suara), Image
Processing/Computer Vision (input/output berupa gambar), dst
 Pembagian Jenis Teknologi Artificial Intelligence (3)

o Berdasar prinsip kerja dalam teknologi AI:

o Problem solving agent
o Solution state space sudah terdefinisi, agent bertugas mencari solusi terbaik dari
solution state space tersebut
o Diselesaikan menggunakan searching algorithm
o Knowledge based agent
o Solution state space belum terdefinisi (non deterministic)
o Agent mencari solusi berdasar knowledge yang dimiliki dimana knowledge dapat
berasal dari expert/sumber informasi atau berdasar knowledge yang dipelajari
(learning agent) dari data
 Problem Solving Agent

o Agent design:
o formulate problem 🡪 search solution 🡪 Agent knows world dynamics
execute World states, actions
o Task Environment: Remember PEAS [when agent doesn’t know 🡪 learning]
o Problem: satisfy goal (goal state) World state is finite, small enough to enumerate
o Agent task: find out which sequence of actions [ when state is infinite 🡪 logic]
World is deterministic
will get it to a goal state [when non-deterministic 🡪 uncertainty]
o 4 components of a problem: initial state, Agent knows current state
operator/successor function, goal test, path [when agent doesn’t know 🡪 logic, uncertainty]
cost Utility for a sequence of states is a sum over path
o Searching: process of looking for sequence of action
o Solution: sequence of action to goal state
Contoh Persoalan pada Problem Solving Agent

O
71 151 F
S 99
Z
75 211
A 140 9
R B
0 97 P
120
118 101
D
75
T M 146 138
111 L 70
C

Map Coloring Route Planning

 Searching Algorithm pada Problem Solving Agent

• UnInformed/Blind Search
• Look around, don’t know where to find the right answer
• No additional information beyond that provided in problem definitional
• Example: DFS, BFS, IDS, UCS , DLS
• Informed Search
• Heuristic Search
• Know some information that sometimes helpful
• Know whether one non-goal state is “more promising” than another
• Example: Best FS, A*,
• Local Search (for Optimization Problem) 🡪 Beyond Classical Search
• Path to goal is irrelevant
• Use very little memory
• Can find reasonable solutions in large or infinite state spaces for which systematic algorithms are
suitable
• Example: Hill-climbing search, simulated annealing search, GA
Knowledge based Agent
 KBS vs Conventional Program

KBS Conventional Program

ill structured problem well structured problem

(uncertain solution, undefined goal, (exact/certain solution, explicit goal,
unknown operator) explicit operator)

expert determine actions, but execution programmer determines actions and

order by interpreter execution order

problem solving method + domain algorithm + data

knowledge + data
 Pendekatan pada KBS

• Symbolic vs Statistical
• Statistical 🡪 speech recognition, vision
• Why symbolic still needed?
• Explanation
• Planning
• Diagnosis
• Many AI systems are hybrid:
• Watson
• SIRI
 Alur Proses dalam KBS

data
learning
Symbolic based KBS: Knowledge Type

• Declarative knowledge
• Know about something: concepts, facts, objects
• Also called descriptive knowledge
• Procedural knowledge
• Knowing how to do something: rule, strategy, procedure,
agenda
• Also known as imperative knowledge
• Meta knowledge
• Knowledge about other type of knowledge
• Heuristic knowledge
• Representing knowledge in a field/subject
• Rules of thumb based on previous experience, good to
work but not guaranteed
• Structural knowledge
• Describe relationship between concepts such as kind-of,
part-of, group of something
Symbolic based KBS: Knowledge Representation Technique

Production Rules
• rule as condition and action pair
• forward & backward chaining

Logical Representation
• propositional logic, first order logic, default logic, etc

Semantic Networks
• knowledge as a form of graphical networks

Frame Representation
• As structure consists of collection of attributes and its values to describe an entity in the world
 Statistical based Knowledge Based System
• Learning Agent
• Changes in the system that are adaptive in the sense that they enable the
system to do the task or tasks drawn from the same population more
efficiently and more effectively the next time
• Why need learning?
• Learning is essential for unknown environment
• i.e., when designer lacks omniscience, agent doesn’t know world
dynamic
• Learning is useful as a system construction method
• i.e., expose the agent to reality rather than trying to write it down
• Learning modifies the agent’s decision mechanism to improve performance
• Learning from observations, feedback for improving the agent’s ability
to act in the future
Taxi Driver as Learning Agent
• Brake decision
• Whether to push the brake or not based on the
current state
• Feedback: instructor shouts
• Buses recognition
• Input: image from camera
• Output: whether it is a bus or not
• Feedback: labeling bus images
 Learning Type

Unsupervised Learning (no feedback)

• Given set of examples without label, detect

potentially useful clusters of input examples,
e.g: customer clustering

Supervised Learning

• Given set of examples (input-output pairs),

learns a function that maps from input to output,
e.g: object classification

Reinforcement Learning

• Agent learns from a series of reinforcements

(rewards or punishments)

Sumber gambar: medium.com

Perkembangan Teknologi AI
Why Now?

● Computing Hardware
○ GPU,TPU, etc
○ Cloud

● Algorithm
○ Deep Learning

● Data availability
○ IoT device, Social Media
Penerapan AI: Monitoring
 Penerapan AI: Verification & Identification
BENEFITS

● Better User Experience (UX)

○ Personalized; Greater ● Reduce Cost Halo
Accuracy; Convenient; ● Accessibility: Disabled person
Frictionless; Fast; (blind, can’t type)
Automated
● Improved Security

USE CASE SCENARIOS

Face Recognition
● Electronic-Know Your Customer
(eKYC)
●
●
Verify Login
Verify Transactions
Voice Biometrics
● “Kependudukan & Pencatatan Sipil” ● Customer Service Conversational
(“Dukcapil”) Analytics

Faster Authentication Accurate & Secure Wide Application

Up to 10 times faster than traditional Using state-of-the-art AI technology to achieve Applicable to banking, financial services, 38
authentication methods* real-time result with guaranteed high accuracy insurance & healthcare, and other industries

Sumber: prosa.ai
Penerapan AI: Document Digitization
Go paperless. Digitize and secure your past, present and
future

BENEFITS

● Reduce Clerical Work ● Automated

● Faster and Accurate Data Input ● Improve User Experience
● Reduce Operational Costs ● Save time
● Convenience ● Save money

USE CASE SCENARIOS

● Electronic-Know Your Customer

(eKYC)
eKTP, NPWP, Passport, ITAS, ● Digitize Documents and Forms
ITAP ● Digitize Name Cards
● Digitize Invoice and Purchase ● Digitize Video/Speech
Speech Orders Recording
Voice ● Credit Card, Birth Certificate,
Transcription
Recording etc
Recognition
process

39
Penerapan AI: Meeting analytics

40
Sumber: prosa.ai
 Tantangan dalam Perkembangan AI

o Regulasi
o Pengaturan etika dan pemakaian AI yang lebih bertanggung-jawab
o Privasi
o Terkait dengan penggunaan data yang dipakai untuk pembangunan model AI
o Kurangnya penjelasan
o Bagaimana model AI sampai pada suatu keputusan/kesimpulan tertentu (terkait
dengan akuntabilitas dan trust)
o Ketersediaan data
o Sejauh mana data yang dipakai cukup representative dan tidak bias
o Kurangnya Talenta
Referensi Video
● Introduction to AI
○ https://www.youtube.com/watch?v=s9vDgPotU-4
○ https://www.youtube.com/watch?v=wfmM5-d0Zh0
○ https://www.youtube.com/watch?v=eUpRwSrwbHY
○ https://www.youtube.com/watch?v=XfEOoAlArXw
○ https://www.youtube.com/watch?v=uyWHthN3Q9c
● Intelligent Agent
○ https://www.youtube.com/watch?v=XqAUPrLu8_s
○ https://www.youtube.com/watch?v=ehXgvsl8i_I
○ https://www.youtube.com/watch?v=NqeVTW4DUuU
○ https://www.youtube.com/watch?v=btffOHgYsBc
○ https://www.youtube.com/watch?v=d2608-UCcR8
○ https://www.youtube.com/watch?v=Spia43l493c
● Introduction to Knowledge based Agent
○ https://www.youtube.com/watch?v=P2DVmc4Zf7I
○ https://www.youtube.com/watch?v=VhKPNctwInw&t=338s
○ https://www.youtube.com/watch?v=7iZWC_NtegM
○ https://www.youtube.com/watch?v=o2alb-eJNqc
Assignment

1. Pelajarilah video yang menjelaskan strategi AI dari sebuah negara dan buatlah rangkuman dari video tsb
2. Rancanglah sebuah Intelligent Agent untuk robot pembersih ruangan dengan mengisikan informasi di
bawah ini:
● Performance indicator: …
● Environment: …
● Actuator: …
● Sensor: ...
1. Rancanglah sebuah ide teknologi AI yang dapat bermanfaat di masa pandemi. Tuliskan deskripsi dari
teknologi tersebut dan tuliskan jenis teknik manakah (searching, knowledge, atau learning) yang
digunakan untuk membangun teknologi tsb, jelaskan alasannya