MACHINE LEARNING

UNIT 1 PART 2
INTRODUCTION

• In 1959, IBM published a paper in the IBM Journal of Research and Development.

• Authored by IBM’s Arthur Samuel, the paper investigated the application of machine learning in the game of
checkers “to verify the fact that a computer can be programmed so that it will learn to play a better game of
checkers than can be played by the person who wrote the program.”

• Although it wasn’t the first published work to use the term “machine learning” per se, Arthur Samuel is
regarded as the first person to coin and define machine learning.
• Samuel’s landmark journal submission, Some Studies in Machine Learning Using the Game of Checkers, introduces
machine learning as a subfield of computer science that gives computers the ability to learn without being explicitly
programmed.

• A key characteristic of machine learning is the concept of self-learning.

• This refers to the application of statistical modeling to detect patterns and improve performance based on data and
empirical information; all without direct programming commands.

• This is what Arthur Samuel described as the ability to learn without being explicitly programmed.

• Samuel didn’t infer that machines may formulate decisions with no upfront programming. On the contrary, machine
learning is heavily dependent on code input. Instead, he observed machines can perform a set task using input data rather
than relying on a direct input command.
• This represents a direct command with a pre-programmed answer, which is typical of most computer
applications.

• Unlike traditional computer programming, though, where outputs or decisions are pre-defined by the
programmer, machine learning uses data as input to build a decision model.
Example –

• Suppose that after analyzing YouTube viewing habits the decision model identifies a significant relationship
among data scientists watching fiction videos.

• A separate model, meanwhile, identifies patterns among the physical traits of baseball players and their
likelihood of winning the season’s Most Valuable Player (MVP) award.

• In the first scenario, the machine analyzed which videos data scientists enjoy watching on YouTube based on
user engagement; measured in likes, subscribes, and repeat viewing.
• In the second scenario, the machine assessed the physical attributes of previous baseball MVPs among other
features such as age and education.
• However, at no stage was the decision model told or programmed to produce those two outcomes.
• By decoding complex patterns in the input data, the model uses machine learning to find connections without
human help.

• This also means that a related dataset gathered from another period of time, with fewer or greater data points,
might lead the model to a slightly different output.
• Another distinct feature of machine learning is the ability to improve predictions based on experience.

• A common example of a self-learning model is a system for detecting spam email messages.

• Following an initial serving of input data, the model learns to flag emails with suspicious subject lines and
body text containing keywords that correlate highly with spam messages flagged by users in the past.

• Indications of spam email may comprise words like dear friend, free, invoice, PayPal, casino, payment,
bankruptcy, and winner.
• However, as the machine is fed more data, it might also find exceptions and incorrect assumptions that render
the model susceptible to bad predictions.

• If there is limited data to reference its decision, the following email subject, for example, might be wrongly
classified as spam:

• “PayPal has received your payment for Casino Royale purchased on eBay.”

• As this is a genuine email sent from a PayPal auto-responder, the spam detection system is lured into
producing a false-positive based on the initial input data.
• Traditional programming is highly susceptible to this problem because the model is rigidly defined according
to pre-set rules.
While data is used to source the self-learning process, more data doesn’t automatically equate to
better decisions; the input data must be relevant to the scope of the model.
 Definition of learning -

• A computer program is said to learn from experience E with respect to some class of tasks T and performance
measure P, if its performance at tasks in T, as measured by P, improves with experience E.
• For example, a computer program that learns to play checkers might improve its performance as measured
by its ability to win at the class of tasks involving playing checkers games, through experience obtained by
playing games against itself.

In general, to have a well-defined learning problem, we must identity these three features:
• the class of tasks T,
• the measure of performance P to be improved,
• and the source of experience E
A checkers learning problem:
• Task T: playing checkers
• Performance measure P: percent of games won against opponents
• Training experience E: playing practice games against itself

Handwriting recognition learning problem

• Task T: Recognising and classifying handwritten words within images
• Performance P: Percent of words correctly classified
• Training experience E: A dataset of handwritten words with given classifications
A robot driving learning problem

• Task T: Driving on highways using vision sensors

• Performance measure P: Average distance travelled before an error
• training experience: A sequence of images and steering commands recorded while observing a human driver

A chess learning problem

• Task T: Playing chess
• Performance measure P: Percent of games won against opponents
• Training experience E: Playing practice games against itself
Definition of ML

• A computer program which learns from experience is called a machine learning program or simply a learning
program. Such a program is sometimes also referred to as a learner.
Relationship between Machine Learning and Human Learning

• Human learning is a natural process that occurs throughout our lives.

• It involves acquiring new knowledge, skills, and behaviors through experience, observation, and instruction.
• Unlike machines, humans have the ability to reason, think critically, and adapt their knowledge to different
contexts.
• The human brain is a complex organ that allows us to process information through various senses, make
connections, and form memories.
Humans learn from variety of sources at different stages of life…

• Biological Learning:
How does the baby learn to cry when she is born, how does she fall upside down after few months, how does
she start sitting, crawling, walking and running? These activities are all biologically and genetically driven. No
one teaches the baby, nor is the baby putting conscious effort to learn.

• Formal Education:
Once the baby starts crawling, her mother starts teaching how to speak few words, which is the precursor to
formal education led by a teacher at play school, primary school, high school, collage, university etc.
• Social Learning:
Humans learn a lot of things from the environment they live in, from friends, relatives, news channels, books,
entertainment, and at times simply from their own experiences etc.

• Cultural Learning:
Based on the culture, community, region we belong to, some beliefs are inherited. We simply follow these
customs/practices/beliefs religiously.
Comparison of Machine Learning and Human Learning

1. Data Processing

• Machine learning algorithms excel at processing and analyzing vast amounts of data quickly
• However, machines rely solely on the data they are trained on and might struggle with new or unexpected
situations where there is a lack of training data. On the other hand, humans have the ability to generalize their
learning to new situations.
2. Contextual Understanding

• Machine learning algorithms lack contextual understanding, as they rely solely on the patterns found in the
data they were trained on. They struggle to understand nuance, sarcasm, or implicit meanings in language or
visual information.
• Humans, on the other hand, possess a deep understanding of context. This contextual understanding allows
humans to engage in complex tasks, such as negotiation, persuasion, and empathy, which currently elude
machines.
3. Adaptability and Generalization

• Machine learning algorithms are highly specialized and typically excel at solving specific tasks. Once they
are trained on a specific dataset, their performance tends to decline when applied to new or different tasks.
Machines lack the flexibility and adaptability that humans possess.
• Human learning, on the other hand, is highly adaptive and can be applied to a wide range of situations. We
can transfer knowledge from one domain to another.
Applications of machine learning