MACHINE

LEARNING
Dr. Supeno Mardi Susiki Nugroho, ST., MT.
 Kelas Logistik dan Jadwal
• 36 Pertemuan
• Software yang digunakan
– Python 3
– TensorFlow (TF) + Keras
• Final Project + Presentasi
 Daftar isi
• Cara Manusia Belajar
• Machine Learning
• Lingkaran Sense, Think, Act: Sistem Robotic
• Supervised learning
• Unsupervised learning
• Pilihan algorithma Machine Learning
• Agen Pintar
• Dunia Robot
 CARA MANUSIA BELAJAR
Bayi belajar dengan cara apa?
• Dari lahir sampai 6 bulan
– Able to recognize the faces of their parents
– Able to discern discrete objects from backgrounds
– Able to tell apart voices
• Dalam setahun
– Developed an intuition for natural physics
– Track objects, even there were occlusion
– Associate sounds with specific meaning
• Batita
– Sophisticated understanding of grammar
– Thousands of words in their vocabularies
 CARA MANUSIA BELAJAR

Prateek Joshi, AI with python

Mengapa perlu AI untuk otomasi macam keperluan
• Kehidupan di dunia:
– We deal with huge and insurmountable amounts of data. The human
brain can't keep track of so much data.
– Data originates from multiple sources simultaneously.
– The data is unorganized and chaotic.
– Knowledge derived from this data has to be updated constantly
because the data itself keeps changing.
– The sensing and actuation has to happen in real time with high
precision.
 Lingkaran Sense, Think, Act
• Sense Sense

– Gather input sensor changes

– Update state with new values Think

• Think
– Poll each decision tree Act
• Act
– Execute any changes to
actions
Arsitektur Sistem Robotic

• Data acquisition
Think • Locomotion
• Filtering • Navigation • Kinematics
• Perception • Localization • Motor control
• Decision making
Sense
Act
 Intelligent machines
• We have dreamed of building intelligent machines with brain like ours
– Domestic service robots = enable to clean home
– Unmanned vehicles = can drive by themselves
Keterbatasan traditional computer programs

• Traditional computer programs are designed to be very good at:

– Performing arithmetic really fast
– Explicitly following a list of instructions
• If we want to solve “interesting” problems such as recognizing human’s handwriting
– What kind of rules could we use?

• Is this a “six” or “zero”?

• How can the computer distinguish
it?

• How about the others?

• How many rules do we have
to write?
MNIST handwritten digit dataset
 Mekanik dari machine learning
• Learn from what we do!
– Our parents do not teach us rules / formula
– We learned from examples  Our moms did that!
• So, machines should have a computational model:
– Doesn’t use a massive list of rules
– Able to evaluate examples
– Can be modified using “small” set of instructions
 Contoh: linear problem
~Solved by a linear perceptron~
• Assume that our model is a function ℎ(𝒙, 𝜃)
– The input 𝒙 = 𝑥1 , 𝑥2 𝑇 is an example expressed in vector form
– 𝑥1 : the number of hours of sleep we got
– 𝑥2 : the number of hours we spent studying

– The input 𝜃 is a vector of the parameters that our model uses

• Machine learning tries to perfect the values of these
parameters as it is exposed to more and more examples

𝒙 = x1 x2 T , 𝜃 = 𝜃0 𝜃1 𝜃2 𝑇
𝜃1
−1 𝑖𝑓 𝒙𝑇 ∙ + 𝜃0 < 0
𝜃2
ℎ(𝒙, 𝜃) =
𝜃
1 𝑖𝑓 𝒙𝑇 ∙ 1 + 𝜃0 ≥ 0
𝜃2
−1 𝑖𝑓 3𝑥1 + 4𝑥2 − 24 < 0
ℎ(𝒙, 𝜃) = ቊ
1 𝑖𝑓 3𝑥1 + 4𝑥2 − 24 ≥ 0
 Contoh: non linear problem
• In general, to solve a problem: finding an “optimal” parameters
– This job is done by an “optimizer” by e.g. minimizing the error
• We need to think another way to solve these!
 The neuron
• The foundation unit of the human brain
– In a tiny piece of brain (about the size of grain of rice)

• 10,000 neurons
• Each of which forms an average of 6,000 connections with other neurons
– They communicate each others by sending and receiving signals from the other
 Apa itu machine learning?
• Machine learning teaches computers to do what comes
naturally to humans and animals: learn from experience
• Machine learning algorithms use computational methods to
“learn” information directly from data without relying on a
predetermined equation as a model
• The algorithms adaptively improve their performance as the
number of samples available for learning increases
 More data, more questions, better answers
• Machine learning algorithms find natural patterns in data that
generate insight and help you make better decisions and
predictions.
• They are used every day to make critical decisions in medical
diagnosis, stock trading, energy load forecasting, and more.
• Media sites rely on machine learning to sift through millions of
options to give you song or movie recommendations.
• Retailers use it to gain insight into their customers’ purchasing
behavior.
 Aplikasi dari Machine Learning
• Computational finance, for credit scoring and algorithmic
trading
• Image processing and computer vision, for face recognition,
motion detection, and object detection
• Computational biology, for tumor detection, drug discovery,
and DNA sequencing
• Energy production, for price and load forecasting
• Automotive, aerospace, and manufacturing, for predictive
maintenance
 Bagaimana machine learning bekerja
• Machine learning uses two types of techniques:
– Supervised learning, which trains a model on known input and
output data so that it can predict future outputs
– Unsupervised learning, which finds hidden patterns or intrinsic
structures in input data
 Supervised learning
• The aim of supervised machine learning is to build a model
that makes predictions based on evidence in the presence of
uncertainty.
– A supervised learning algorithm takes a known set of input data and
known responses to the data (output) and trains a model to generate
reasonable predictions for the response to new data
• Supervised learning uses classification and regression
techniques to develop predictive models
 Classification techniques
• Predict discrete responses
– For example, whether an email is genuine or spam, or whether a tumor is cancerous or benign
• Classification models classify input data into categories
• Typical applications include medical imaging, speech recognition, and credit scoring
 Regression techniques
• Predict continuous responses
– For example, changes in temperature or fluctuations in power
demand.
• Typical applications include electricity load forecasting and
algorithmic trading
 Unsupervised learning
• Finds hidden patterns or intrinsic structures in data
– It is used to draw inferences from datasets consisting of input data
without labeled responses
• Clustering is the most common unsupervised learning
technique
– It is used for exploratory data analysis to find hidden patterns or
groupings in data
• Applications for clustering include gene sequence analysis,
market research, and object recognition
 Algoritma Machine Learning
• Choosing the right algorithm can seem overwhelming
– There are dozens of supervised and unsupervised machine learning
algorithms, and each takes a different approach to learning
• There is no best method or one size fits all
– Finding the right algorithm is partly just trial and error
• even highly experienced data scientists can’t tell whether an algorithm will
work without trying it out.
– But algorithm selection also depends on the size and type of data
you’re working with
• The insights you want to get from the data
• How those insights will be used
Pilihan algorithma Machine Learning

KNN

SVM
 When should you use machine
learning?
• Consider using machine learning when:
– Have a complex task or problem involving:

• A large amount of data

• Lots of variables
– But, no existing formula or equation.
– For example, machine learning is a good option if you need to handle situations like these:
Real-World Examples
Real-World Examples
Real-World Examples
 Agen pintar
• What is an agent?
– An entity capable of combining cognition, perception and action in
behaving autonomously, purposively and flexibly in some
environment
• What does an agent do?
– An agent perceives its environment, reasons about its goals, and acts
upon the environment
• What does PEAS stand for (description of an agent)?
– Performance
Russel measure, Environment, Actuators, Sensors
Norvig, AI Modern Approach

• What are we about to do right now? 30

 Generic Agent Program
• Implementing f : P*  A …or… f (P*) = A
– Lookup table?
– Learning? Knowledge, past percepts, past actions

e.g.,
Table-Driven-Agent
Add percept to percepts
LUT [percepts, table]
NOP

31
 Basic types of agent programs
• Simple reflex agent
• Model-based reflex agent
• Goal-based agent
• Utility-based agent
• Learning agent

32
 Simple Reflex Agent

• Input/output associations
• Condition-action rule: “If-then” rule (production rule)
– If condition then action (if in a certain state, do this)
– If antecedent then consequent

33
 Simple Reflex Agent

• Simple state-based agent – Classify the current percept into a known state,
then apply the rule for that state

34
 Model-Based Reflex Agent

• Internal state – keeps track of the world,

models the world

35
 Model-Based Reflex Agent

• State-based agent – Given the current state, classify the current percept
into a known state, then apply the rule for that state

36
 Goal-Based Agent

• Goal: immediate, or long sequence of actions?

– Search and planning – finding action sequences that
achieve the agent’s goals

37
 Utility-Based Agent

• “There are many ways to skin a cat”

• Utility function: Specifies degree of usefulness (happiness)
1
U 
– Maps a state onto a real number cost

38
Learning Agent

39
 Daftar isi
• Cara Manusia Belajar
• Machine Learning
• Lingkaran Sense, Think, Act: Sistem Robotic
• Supervised learning
• Unsupervised learning
• Pilihan algorithma Machine Learning
• Agen Pintar
• Dunia Robot
 A Robot Platform for every Age and Application

NI LabVIEW

LEGO ® MINDSTORMS® NXT

powered by LabVIEW

LEGO ® WeDo®
powered by LabVIEW

NI
WorldClass
 Types ofRobots
• Fixed-base robots

• Mobile robots (tele-

operated)

• Autonomous vehicles and systems

 Vision GuidedMotion
Application
Pick and place parts of different sizes and orientations

Challenges
Multiple domains of expertise need to be addressed

Key benefits
Integration of Robot Control and Vision in a single development
environment
Underwater DredgingSystem: Remote Control
Nexans Spider
Application
Level the seabed floor to lay a pipeline from the Orman Lange natural gas field to
the UK. Develop HMI, remote-control interface, heave compensation and power
controls

Challenges
1000 meters below sea level, strong currents,
uneven terrain, constrained time and budget
Key benefits
Rapid development software across multiple computing platforms, rugged
embedded platform
 Autonomous Vehicle forDARPA UrbanChallenge
• TORC Technologies and Virginia Tech “Odin” Vehicle
• Application
• Autonomously navigate 60 mile urban environment

• Challenges
• Follow all traffic laws
• Safely navigate around obstacles/barriers Safely interact with other UGV and manned vehicles at stop signs during passing,
parking

• Key benefits
• Rapid development platform
• Domain expertise vs. computer science focus Common tool across Windows, Linux, VxWorks, FPGA
• Fast design iterations
• Infrastructure of support for NI and 3rd party I/O hardware: unrivaled by any text-based language