Scribd
Upload
7 views30 pages

The Machine Learning Landscape

The document provides an overview of machine learning, including its definitions, types, and applications. It covers fundamental concepts such as supervised, unsupervised, semi-supervised, and reinforcement learning, as well as challenges faced in machine learning. Additionally, it highlights the importance of data quality and feature selection in building effective machine learning models.

Uploaded by

awetbrhanu122119
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
7 views30 pages

The Machine Learning Landscape

The document provides an overview of machine learning, including its definitions, types, and applications. It covers fundamental concepts such as supervised, unsupervised, semi-supervised, and reinforcement learning, as well as challenges faced in machine learning. Additionally, it highlights the importance of data quality and feature selection in building effective machine learning models.

Uploaded by

awetbrhanu122119
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 30

Introduction

Machine Learning
Outlines

 Part I. The Fundamentals of Machine Learning

Chap-1. Introduction
Chap-2. Supervised Learning
Chap-3. Unsupervised Learning
Chap-4. Reinforcement Learning

 Part II. Deep Learning


Chap-5. Introduction to Artificial Neural Networks
Chap-6. Sequence Modeling: Recurrent and
Recursive Nets
Chapter-1
Cont.

Artificial
Intelligence

Machine
Learning

Machine Learning
Deep Data Science
Learning
Cont.
 Artificial intelligence (AI): is computer software that
mimics human cognitive abilities in order to perform
complex tasks that historically could only be done by
humans, such as decision making, data analysis, and
language translation.
 Machine learning (ML): Machine learning is a subset of AI
in which algorithms are trained on data sets to become
machine learning models capable of performing specific
tasks.
 Deep learning: Deep learning is a subset of ML, in which
artificial neural networks (ANNs) that mimic the human
brain are used to perform more complex reasoning tasks
without human intervention.
 Data Science: Is all about building complex predictive and
machine learning models to solving business problems.
Cont.
Cont.
 What Is Machine Learning?
 “Learning is any process by which a system
improves performance from experience.”
- Herbert Simon
Machine Learning is the science (and art) of
programming computers so they can learn
from data.
Here is a slightly more general definition:
 Machine Learning is the field of study that

gives computers the ability to learn without


being explicitly programmed. [ Arthur
Samuel, 1959]
Cont.
 Definition by Tom Mitchell (1998):
 Machine Learning is the study of algorithms
that
 improve their performance P
 at some task T

 with experience E.

A well-defined learning task is given by <P,


T, E>.
 It is a research field at the intersection of
statistics, artificial intelligence, and
computer science and is also known as
Why Machine Learning?
Cont.
 Consider how you would write a spam filter using
traditional programming techniques
Cont.
Cont.
Cont.
 Applying ML techniques to dig into large amounts
of data can help discover patterns that were not
immediately apparent. This is called data mining.
Cont.
 To summarize, Machine Learning is great for:
 Problems for which existing solutions require a lot
of hand-tuning or long lists of rules: one Machine
Learning algorithm can often simplify code and
perform better.
 Complex problems for which there is no good
solution at all using a traditional approach: the best
Machine Learning techniques can find a solution.
 Fluctuating environments: a Machine Learning
system can adapt to new data.
 Getting insights about complex problems and large
amounts of data.
Types of Machine Learning Systems

 Whether or not they are trained with


human supervision (supervised,
unsupervised, semi-supervised, and
Reinforcement Learning)
 Whether or not they can learn
incrementally on the fly (online versus
batch learning)
 Whether they work by simply
comparing new data points to known
data points, or instead detect patterns in
the training data and build a predictive
Con.
1) Supervised/Unsupervised Learning
Supervised learning (classification vs.
regression)
 In supervised learning, the training data you feed
to the algorithm includes the desired solutions,
called labels
Con.
 Most important supervised learning
algorithms
 k-Nearest Neighbors

 Linear Regression

 Logistic Regression

 Support Vector Machines (SVMs)

 Decision Trees and Random Forests

 Neural networks
Con.
 Unsupervised learning
 In unsupervised learning, as you might guess, the
training data is unlabeled
 The system tries to learn without a teacher.
Con.
 Most important unsupervised learning algorithms
 Clustering

—k-Means
—Hierarchical Cluster Analysis (HCA)
—Expectation Maximization
 Visualization and dimensionality reduction
—Principal Component Analysis (PCA)
—Kernel PCA
—Locally-Linear Embedding (LLE)
—t-distributed Stochastic Neighbor
Embedding
 Association rule learning
—Apriori
Con.

 Semi-supervised learning
 Some algorithms can deal with partially labeled
training data, usually a lot of unlabeled data and a
little bit of labeled data. This is called semi-
supervised learning
Con.
 Reinforcement Learning
 The learning system, called an agent in this context,
can observe the environment, select and perform
actions, and get rewards in return
Con.
2) Batch and Online Learning
 Batch learning
In batch learning, the system is incapable of
learning incrementally
It must be trained using all the available data, then
evaluated on separate test data.
 This will generally take a lot of time and computing
resources
 CPU, memory space,

 disk space, disk I/O, network I/O, so it is typically

done offline.
 First the system is trained, and then it is launched
into production and runs without learning anymore;
it just applies what it has learned.

Con.

 Online learning
 You train the system incrementally by feeding it data
instances sequentially, either individually or by small
groups called mini-batches.
 Each learning step is fast and cheap, so the system
can learn about new data on the fly, as it arrives.
Con.

3) Instance-Based Versus Model-Based Learning


 Given a number of training examples, the
system needs to be able to generalize to
examples it has never seen before.
 Having a good performance measure on the
training data is good, but insufficient; the true
goal is to perform well on new instances.
 There are two main approaches to
generalization
 Instance-based learning and
 Model-based learning
Cont.
 Supervised (inductive) learning
– Given: training data + desired outputs
(labels)
 Unsupervised learning
– Given: training data (without desired
outputs)
 Semi-supervised learning
– Given: training data + a few desired
outputs
 Reinforcement learning
Machine Learning Challenges
 Insufficient Quantity of Training Data
 Non-representative Training Data
 The training data must be representative of the new
cases you want to generalize to.
 Poor-Quality Data
 The training data must be free of errors, outliers,
and noise
 Irrelevant Features
 The system will only be capable of learning if the training data
contains enough relevant features.
 feature engineering
 Feature selection: selecting the most useful features to train
on among existing features.
 Feature extraction: combining existing features to produce a
more useful one
Con.
 Over-fitting the Training Data
 A model that works well on the training set but is not
able to generalize to new data
 Under-fitting the Training Data
Cont.
Cont.
Testing and Validating
Sample Applications
 Recognizing patterns:
– Facial identities or facial expressions
– Handwritten or spoken words
– Medical images
 Generating patterns:
– Generating images or motion sequences
 Recognizing anomalies:
– Unusual credit card transactions
– Unusual patterns of sensor readings in a
nuclear power plant
 Prediction:
– Future stock prices or currency exchange
rates

You might also like