34,79 €
Mehr erfahren.
- Herausgeber: Packt Publishing
- Kategorie: Lebensstil
- Sprache: Englisch
Most of us have heard about the term Machine Learning, but surprisingly the question frequently asked by developers across the globe is, “How do I get started in Machine Learning?”. One reason could be attributed to the vastness of the subject area because people often get overwhelmed by the abstractness of ML and terms such as regression, supervised learning, probability density function, and so on. This book is a systematic guide teaching you how to implement various Machine Learning techniques and their day-to-day application and development.
You will start with the very basics of data and mathematical models in easy-to-follow language that you are familiar with; you will feel at home while implementing the examples. The book will introduce you to various libraries and frameworks used in the world of Machine Learning, and then, without wasting any time, you will get to the point and implement Regression, Clustering, classification, Neural networks, and more with fun examples. As you get to grips with the techniques, you’ll learn to implement those concepts to solve real-world scenarios for ML applications such as image analysis, Natural Language processing, and anomaly detections of time series data.
By the end of the book, you will have learned various ML techniques to develop more efficient and intelligent applications.
Das E-Book können Sie in Legimi-Apps oder einer beliebigen App lesen, die das folgende Format unterstützen:
Veröffentlichungsjahr: 2017
Ähnliche
BIRMINGHAM - MUMBAI
Machine Learning for Developers
Copyright © 2017 Packt Publishing
All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, without the prior written permission of the publisher, except in the case of brief quotations embedded in critical articles or reviews.
Every effort has been made in the preparation of this book to ensure the accuracy of the information presented. However, the information contained in this book is sold without warranty, either express or implied. Neither the author, nor Packt Publishing, and its dealers and distributors will be held liable for any damages caused or alleged to be caused directly or indirectly by this book.
Packt Publishing has endeavored to provide trademark information about all of the companies and products mentioned in this book by the appropriate use of capitals. However, Packt Publishing cannot guarantee the accuracy of this information.
First published: October 2017
Production reference: 1241017
ISBN 978-1-78646-987-8
www.packtpub.com
Credits
Author
Rodolfo Bonnin
Copy Editor
Safis Editing
Reviewers
Doug Ortiz
Md. Mahmudul Hasan
Project Coordinator
Nidhi Joshi
Commissioning Editor
Veena Pagare
Proofreader
Safis Editing
Acquisition Editor
Vinay Argekar
Indexer
Francy Puthiry
Content Development Editor
Aishwarya Pandere
Graphics
Tania Dutta
Technical Editor
Karan Thakkar
Production Coordinator
Shraddha Falebhai
Foreword
Several technological drivers that arose in the last decade that I made big data possible, irreversibly reshaping the world completely. Among these, machine learning plays a singular role, since it provides the main functionalities required for data analysis, mining, knowledge discovery, and many other features that provide actionable autonomous intelligence in a way invisible but pervasive to most of the systems we use in our daily lives. Although not new, the formalisms and methods within machine learning have quickly evolved, driven by the growing demands of e-commerce, social networks, internet-related services and products, and similar enterprises centered on online business.Breakthroughs in machine learning have been fueled by other technological innovations sprouted and matured within the Hadoop ecosystem, including horizontally scalable computational resources and superior warehousing capabilities that have made the real-time analysis of huge datasets feasible. At the same time, community supported initiatives around the Python programming language have made the use and evolution of sophisticated analysis libraries widspread, giving rise to a remarkable amount of knowledge and experience, at the same time fast and easy to deploy and to put into production.Within machine learning, neural networks play a singular role nowadays. Perhaps the first artificial intelligence paradigm to be proposed more than 70 years ago, neural networks have experienced several cycles of being abandoned by the community only to be rediscovered some years later. This was likely due to the lack of computational power to perform really complex analysis adequately, together with the burdensome task of assembling, training, and testing different topologies by trial and error. This has changed dramatically in recent years, mostly due to the availability of cloud computing, GPUs, and programming libraries that allow the set up of networks with simple scripts. Today, a network with hundreds of millions of degrees of freedom can be assembled in minutes, trained in hours, and put into production in a few days, (obviously, if you know the right technologies to so). This is one of the reasons why most of the radical advancements in computer vision, language understanding, and pattern recognition in general are being driven specifically by different flavors of neural networks that have been proposed recently.
This exponentially growing set of knowledge, techniques, and programming libraries makes most classical texts on the subject obsolete, at least for the deployment of fast and practical applications. For this reason, a book like this can be celebrated as a quick and to-the-point text that provides all the materials required to successfully implement and understand a machine learning application in a single reading. In this book, you will find:
The fundamentals of machine learning tasks (classification, clustering, regression, and data reduction), together with a quick, yet comprehensive introduction to the mathematical and statistical foundations of the subject.
A more detailed presentation of Neural Networks as a learning model, together with basics of the training algorithms, convergence crite0ria, and the evaluation of results.
An introduction the most advanced learning models using more elaborate networks, including convolutional, recurrent, and adversarial networks. Each of the models is analyzed thoroughly, both in theoretical and in practical considerations.
A comprehensive guide to open source software that, together with the previous material, allows the reader to put the concepts into practice very quickly.
This book is highly recommended for practitioners in academia who feel their expertise is becoming outdated, for developers who need to deploy sophisticated machine learning features in business applications, and for anyone willing to gain a broad and practical understanding of machine learning. The author transmits his vast experience in the subject in a very clear and systematic manner, making the book easy to follow and put into practice.
Claudio Delrieux Full Professor at Electric and Computer Engineering Department - Universidad Nacional del Sur Fellow - National Research and Technology Council of Argentina Chair - Imaging Sciences Laboratory
About the Author
Rodolfo Bonnin is a systems engineer and Ph.D. student at Universidad Tecnológica Nacional, Argentina. He has also pursued parallel programming and image understanding postgraduate courses at Universität Stuttgart, Germany.
He has been doing research on high-performance computing since 2005 and began studying and implementing convolutional neural networks in 2008, writing a CPU- and GPU-supporting neural network feedforward stage. More recently he's been working in the field of fraud pattern detection with Neural Networks and is currently working on signal classification using machine learning techniques.
He is also the author of Building Machine Learning Projects with Tensorflow, by Packt Publishing.
About the Reviewers
Doug Ortiz is a Senior Big Data Architect at ByteCubed who has been architecting, developing, and integrating enterprise solutions throughout his whole career. Organizations that leverage his skillset have been able to rediscover and reuse their underutilized data via existing and emerging technologies such as Amazon Web Services, Microsoft Azure, Google Cloud, Microsoft BI stack, Hadoop, Spark, NoSQL databases, SharePoint along with related toolsets and technologies.
He is also, the founder of Illustris, LLC and can be reached at [email protected]
Interesting aspects of his profession include:
Experience integrating multiple platforms and products
Has Big Data, Data Science Certifications, R, and Python Certifications
Helping organizations gain a deeper understanding and value of their current investments in data and existing resources turning them into useful sources of information
Improving, salvaging and architecting projects by utilizing unique and innovative techniques
Reviews regularly books on the topics of Amazon Web Services, Data Science, Machine Learning, R and Cloud Technologies
Hobbies:
Yoga
Scuba diving
Md. Mahmudul Hasan is currently serving as a Ph.D. researcher at Anglia Ruskin IT Research Institute (ARITI) at Anglia Ruskin University, UK. He worked as a senior lecturer in the department of CSE at Daffodil International University. He graduated from University of Essex, UK and specialized in games and mobile apps development in different platforms.
He has 6+ years of experiences in ICT industry including development to commercialize of small and large software. He is currently serving himself as a country Director and Chair of International Game Developers Association (IGDA) Bangladesh Chapter.
Mahmud's research interests include machine learning, data science, decision support system, and personalized learning through games and gamification. He has significant publications in peer-reviewed journals and conferences.
www.PacktPub.com
For support files and downloads related to your book, please visit www.PacktPub.com.
Did you know that Packt offers eBook versions of every book published, with PDF and ePub files available? You can upgrade to the eBook version at www.PacktPub.com and as a print book customer, you are entitled to a discount on the eBook copy. Get in touch with us at [email protected] for more details.
At www.PacktPub.com, you can also read a collection of free technical articles, sign up for a range of free newsletters and receive exclusive discounts and offers on Packt books and eBooks.
https://www.packtpub.com/mapt
Get the most in-demand software skills with Mapt. Mapt gives you full access to all Packt books and video courses, as well as industry-leading tools to help you plan your personal development and advance your career.
Why subscribe?
Fully searchable across every book published by Packt
Copy and paste, print, and bookmark content
On demand and accessible via a web browser
Customer Feedback
Thanks for purchasing this Packt book. At Packt, quality is at the heart of our editorial process. To help us improve, please leave us an honest review on this book's Amazon page at https://www.amazon.com/dp/1786466961.
If you'd like to join our team of regular reviewers, you can e-mail us at [email protected]. We award our regular reviewers with free eBooks and videos in exchange for their valuable feedback. Help us be relentless in improving our products!
Table of Contents
Preface
What this book covers
What you need for this book
Who this book is for
Conventions
Reader feedback
Customer support
Downloading the example code
Errata
Piracy
Questions
Introduction - Machine Learning and Statistical Science
Machine learning in the bigger picture
Types of machine learning
Grades of supervision
Supervised learning strategies - regression versus classification
Unsupervised problem solving–clustering
Tools of the trade–programming language and libraries
The Python language
The NumPy library
The matplotlib library
What's matplotlib?
Pandas
SciPy
Jupyter notebook
Basic mathematical concepts
Statistics - the basic pillar of modeling uncertainty
Descriptive statistics - main operations
Mean
Variance
Standard deviation
Probability and random variables
Events
Probability
Random variables and distributions
Useful probability distributions
Bernoulli distributions
Uniform distribution
Normal distribution
Logistic distribution
Statistical measures for probability functions
Skewness
Kurtosis
Differential calculus elements
Preliminary knowledge
In search of changes–derivatives
Sliding on the slope
Chain rule
Partial derivatives
Summary
The Learning Process
Understanding the problem
Dataset definition and retrieval
The ETL process
Loading datasets and doing exploratory analysis with SciPy and pandas
Working interactively with IPython
Working on 2D data
Feature engineering
Imputation of missing data
One hot encoding
Dataset preprocessing
Normalization and feature scaling
Normalization or standardization
Model definition
Asking ourselves the right questions
Loss function definition
Model fitting and evaluation
Dataset partitioning
Common training terms – iteration, batch, and epoch
Types of training – online and batch processing
Parameter initialization
Model implementation and results interpretation
Regression metrics
Mean absolute error
Median absolute error
Mean squared error
Classification metrics
Accuracy
Precision score, recall, and F-measure
Confusion matrix
Clustering quality measurements
Silhouette coefficient
Homogeneity, completeness, and V-measure
Summary
References
Clustering
Grouping as a human activity
Automating the clustering process
Finding a common center - K-means
Pros and cons of K-means
K-means algorithm breakdown
K-means implementations
Nearest neighbors
Mechanics of K-NN
Pros and cons of K-NN
K-NN sample implementation
Going beyond the basics
The Elbow method
Summary
References
Linear and Logistic Regression
Regression analysis
Applications of regression
Quantitative versus qualitative variables
Linear regression
Determination of the cost function
The many ways of minimizing errors
Analytical approach
Pros and cons of the analytical approach
Covariance/correlation method
Covariance
Correlation
Searching for the slope and intercept with covariance and correlation
Gradient descent
Some intuitive background
The gradient descent loop
Formalizing our concepts
Expressing recursion as a process
Going practical – new tools for new methods
Useful diagrams for variable explorations – pairplot
Correlation plot
Data exploration and linear regression in practice
The Iris dataset
Getting an intuitive idea with Seaborn pairplot
Creating the prediction function
Defining the error function
Correlation fit
Polynomial regression and an introduction to underfitting and overfitting
Linear regression with gradient descent in practice
Logistic regression
Problem domain of linear regression and logistic regression
Logistic function predecessor – the logit functions
Link function
Logit function
Logit function properties
The importance of the logit inverse
The sigmoid or logistic function
Properties of the logistic function
Multiclass application – softmax regression
Practical example – cardiac disease modeling with logistic regression
The CHDAGE dataset
Dataset format
Summary
References
Neural Networks
History of neural models
The perceptron model
Improving our predictions – the ADALINE algorithm
Similarities and differences between a perceptron and ADALINE
Limitations of early models
Single and multilayer perceptrons
MLP origins
The feedforward mechanism
The chosen optimization algorithm – backpropagation
Types of problem to be tackled
Implementing a simple function with a single-layer perceptron
Defining and graphing transfer function types
Representing and understanding the transfer functions
Sigmoid or logistic function
Playing with the sigmoid
Rectified linear unit or ReLU
Linear transfer function
Defining loss functions for neural networks
L1 versus L2 properties
Summary
References
Convolutional Neural Networks
Origin of convolutional neural networks
Getting started with convolution
Continuous convolution
Discrete convolution
Kernels and convolutions
Stride and padding
Implementing the 2D discrete convolution operation in an example
Subsampling operation (pooling)
Improving efficiency with the dropout operation
Advantages of the dropout layers
Deep neural networks
Deep convolutional network architectures through time
Lenet 5
Alexnet
The VGG model
GoogLenet and the Inception model
Batch-normalized inception V2 and V3
Residual Networks (ResNet)
Types of problem solved by deep layers of CNNs
Classification
Detection
Segmentation
Deploying a deep neural network with Keras
Exploring a convolutional model with Quiver
Exploring a convolutional network with Quiver
Implementing transfer learning
References
Summary
Recurrent Neural Networks
Solving problems with order — RNNs
RNN definition
Types of sequence to be modeled
Development of RNN
Training method — backpropagation through time
Main problems of the traditional RNNs — exploding and vanishing gradients
LSTM
The gate and multiplier operation
Part 1 — set values to forget (input gate)
Part 2 — set values to keep
Part 3 — apply changes to cell
Part 4 — output filtered cell state
Univariate time series prediction with energy consumption data
Dataset description and loading
Dataset preprocessing
Summary
References
Recent Models and Developments
GANs
Types of GAN applications
Discriminative and generative models
Reinforcement learning
Markov decision process
Decision elements
Optimizing the Markov process
Basic RL techniques: Q-learning
References
Summary
Software Installation and Configuration
Linux installation
Initial distribution requirements
Installing Anaconda on Linux
pip Linux installation method
Installing the Python 3 interpreter
Installing pip
Installing necessary libraries
macOS X environment installation
Anaconda installation
Installing pip
Installing remaining libraries via pip
Windows installation
Anaconda Windows installation
Summary
Preface
Machine learning is one of the star disciplines at present. Acclaimed by the media as the future of work, it is said to be part of any significant tech investment in recent months, in a world where everything is driven by data and automation. It is used extensively across many fields such as image understanding, robotics, search engines, self-driving cars, and so on and the number of areas of application increases almost daily. In this book we will study the motivations and current techniques of machine learning using code and diagrams as the main conceptual vehicles, omitting outside the fundamental mathematical results.
We will start talking about the fundamental machine learning concepts, its branches, and types of problems. Then, there will be an explanatory chapter about the fundamental mathematical concepts required to grasp upcoming techniques. As we advance through the chapters, models of increasing complexity and sophistication are explained, starting with linear regression, then logistic regression, neural networks and its more recent variants (CNNs, RNNs),concluding with a synthetic introduction to more advanced machine learning techniques, such as GANs and reinforcement learning.
This book is aimed at developers looking to finally grasp what that machine learning hype is all about, and understand the main fundamental concepts, using an algorithmic point of view, along with more formal mathematical definitions.This book implements code concepts in Python, considering the simplicity of its interface, and the fact that Python offers an unmatched set of tools to continue learning from the book’s code. So, familiarity with Python programming would certainly be helpful for playing around with the code, but it should be manageable from programmers experienced in other languages.
You will learn how to make informed decisions about the types of algorithms you need to solve your own machine learning problems, and a knowledge of how those algorithms work to get the best possible results. If you want to understand machine learning in day-to-day, coder-friendly language, and have just the right amount of information to be able to do the jump into the discipline, this book will definitely come to your rescue!
What this book covers
Chapter 1, Introduction - Machine Learning and Statistical Science, covers various introductory concepts in machine learning. It talks about the history, branches and general discipline concepts. It also gives an introduction to the base mathematical concepts needed to understand most of the techniques developed afterward.
Chapter 2, The Learning Process, covers all the steps in the workflow of a machine learning process and shows useful tools and concept definitions for all those stages.
Chapter 3, Clustering, covers several techniques for unsupervised learning, specially K-Means, and K-NN clustering.
Chapter 4, Linear and Logistic Regression, covers two pretty different supervised learning algorithms, which go under a similar name: linear regression (which we will use to perform time series predictions), and logistic regression (which we will use for classification purposes).
Chapter 5, Neural Networks, covers one of the basic building blocks of modern machine learning Applications, and ends with the practical step-by-step building of a neural network.
Chapter 6, Convolutional Neural Networks, covers this powerful variation of neural networks, and ends with a practical tour of the internals of a very well known architecture of CNN, called VGG16, in a practical application.
Chapter 7, Recurrent Neural Networks, covers an overview of the RNN concept and a complete depiction of all the stages of the most used architecture, the LSTM. Finally, a practical exercise in time series prediction is shared.
Chapter 8, Recent Models and Developments, covers two upcoming techniques that have engaged huge interest in the field: generative adversarial networks, and the whole reinforcement learning field.
Chapter 9, Software Installation and Configuration, It covers the installation of all the necessary software packages, for three operative systems: Linux, macOS, and Windows.
What you need for this book
This book is focused on machine learning concepts and uses as a Python language (version 3) as a computational tool . We have used Python 3 and the Jupyter Notebook to build our workbooks, which you can edit and play with to better understand the concepts. We focus on how to utilize various Python libraries in the best possible way to build real-world applications. In that spirit, we have tried to keep all the code as friendly and readable as possible. We feel that this will enable our readers to easily understand the code and readily use it in different scenarios.
Who this book is for
This book is for developers/tech enthusiasts who want to understand the basics of machine learning concepts through a computationally-intensive approach. This book should be suited to people who have programmed in any scripting language, but being familiar with Python will be useful to play around with the code. It will also be useful for current data scientists to get back to the basic concepts, and understand them using a novel, hands-on approach.
Conventions
In this book, you will find a number of text styles that distinguish between different kinds of information. Here are some examples of these styles and an explanation of their meaning.
Code words in text, database table names, folder names, filenames, file extensions, path names, dummy URLs, user input, and Twitter handles are shown as follows: "Let's generate a large number of events of a Bernoulli distribution using np and graph the tendency of this distribution."
A block of code is set as follows:
def mean(sampleset): #Definition header for the mean function total=0 for element in sampleset: total=total+element return total/len(sampleset)
New terms and important words are shown in bold. Words that you see on the screen, for example, in menus or dialog boxes, appear in the text like this: "In order to download new modules, we will go toFiles|Settings|Project Name|Project Interpreter."
Reader feedback
Feedback from our readers is always welcome. Let us know what you think about this book-what you liked or disliked. Reader feedback is important to us as it helps us develop titles that you will really get the most out of. To send us general feedback, simply email [email protected], and mention the book's title in the subject of your message. If there is a topic that you have expertise in and you are interested in either writing or contributing to a book, see our author guide at www.packtpub.com/authors.
Customer support
Now that you are the proud owner of a Packt book, we have a number of things to help you to get the most from your purchase.
Downloading the example code
You can download the example code files for this book from your account at http://www.packtpub.com. If you purchased this book elsewhere, you can visit http://www.packtpub.com/support and register to have the files emailed directly to you. You can download the code files by following these steps:
Log in or register to our website using your email address and password.
Hover the mouse pointer on the
SUPPORT
tab at the top.
Click on
Code Downloads & Errata
.
Enter the name of the book in the
Search
box.
Select the book for which you're looking to download the code files.
Choose from the drop-down menu where you purchased this book from.
Click on
Code Download
.
Once the file is downloaded, please make sure that you unzip or extract the folder using the latest version of:
WinRAR / 7-Zip for Windows
Zipeg / iZip / UnRarX for Mac
7-Zip / PeaZip for Linux
The code bundle for the book is also hosted on GitHub at https://github.com/PacktPublishing/Machine-Learning-for-Developers. We also have other code bundles from our rich catalog of books and videos available at https://github.com/PacktPublishing/. Check them out!
Errata
Although we have taken every care to ensure the accuracy of our content, mistakes do happen. If you find a mistake in one of our books-maybe a mistake in the text or the code-we would be grateful if you could report this to us. By doing so, you can save other readers from frustration and help us improve subsequent versions of this book. If you find any errata, please report them by visiting http://www.packtpub.com/submit-errata, selecting your book, clicking on the Errata Submission Form link, and entering the details of your errata. Once your errata are verified, your submission will be accepted and the errata will be uploaded to our website or added to any list of existing errata under the Errata section of that title. To view the previously submitted errata, go to https://www.packtpub.com/books/content/support and enter the name of the book in the search field. The required information will appear under the Errata section.
Piracy
Piracy of copyrighted material on the internet is an ongoing problem across all media. At Packt, we take the protection of our copyright and licenses very seriously. If you come across any illegal copies of our works in any form on the internet, please provide us with the location address or website name immediately so that we can pursue a remedy. Please contact us at [email protected] with a link to the suspected pirated material. We appreciate your help in protecting our authors and our ability to bring you valuable content.
Questions
If you have a problem with any aspect of this book, you can contact us at [email protected], and we will do our best to address the problem.
Introduction - Machine Learning and Statistical Science
Machine learning has definitely been one of the most talked about fields in recent years, and for good reason. Every day new applications and models are discovered, and researchers around the world announce impressive advances in the quality of results on a daily basis.
Each day, many new practitioners decide to take courses and search for introductory materials so they can employ these newly available techniques that will improve their applications. But in many cases, the whole corpus of machine learning,as normally explained in the literature, requires a good understanding of mathematical concepts as a prerequisite, thus imposing a high bar for programmers who typically have good algorithmic skills but are less familiar with higher mathematical concepts.
This first chapter will be a general introduction to the field, covering the main study areas of machine learning, and will offer an overview of the basic statistics, probability, and calculus, accompanied by source code examples in a way that allows you to experiment with the provided formulas and parameters.
In this first chapter, you will learn the following topics:
What is machine learning?
Machine learning areas
Elements of statistics and probability
Elements of calculus
The world around us provides huge amounts of data. At a basic level, we are continually acquiring and learning from text, image, sound, and other types of information surrounding us. The availability of data, then, is the first step in the process of acquiring the skills to perform a task.
A myriad of computing devices around the world collect and store an overwhelming amount of information that is image-, video-, and text-based. So, the raw material for learning is clearly abundant, and it's available in a format that a computer can deal with.
That's the starting point for the rise of the discipline discussed in this book: the study of techniques and methods allowing computers to learn from data without being explicitly programmed.
A more formal definition of machine learning, from Tom Mitchell, is as follows:
This definition is complete, and reinstates the elements that play a role in every machine learning project: the task to perform, the successive experiments, and a clear and appropriate performance measure. In simpler words, we have a program that improves how it performs a task based on experience and guided by a certain criterion.
Machine learning in the bigger picture
Machine learning as a discipline is not an isolated field—it is framed inside a wider domain, Artificial Intelligence (AI). But as you can guess, machine learning didn't appear from the void. As a discipline it has its predecessors, and it has been evolving in stages of increasing complexity in the following four clearly differentiated steps:
The first model of machine learning involved rule-based decisions and a simple level of data-based algorithms that includes in itself, and as a prerequisite, all the possible ramifications and decision rules, implying that all the possible options will be hardcoded into the model beforehand by an expert in the field. This structure was implemented in the majority of applications developed since the first programming languages appeared in 1950. The main data type and function being handled by this kind of algorithm is the Boolean
, as it exclusively dealt with yes or no decisions.
During the second developmental stage of statistical reasoning, we started to let the probabilistic characteristics of the data have a say, in addition to the previous choices set up in advance. This better reflects the fuzzy nature of real-world problems, where outliers are common and where it is more important to take into account the nondeterministic tendencies of the data than the rigid approach of fixed questions. This discipline adds to the mix of mathematical tools elements of
Bayesian probability theory
. Methods pertaining to this category include curve fitting (usually of linear or polynomial), which has the common property of working with numerical data.
The machine learning stage is the realm in which we are going to be working throughout this book, and it involves more complex tasks than the simplest Bayesian elements of the previous stage. The most outstanding feature of machine learning algorithms is that they can generalize models from data but the models are capable of generating their own feature selectors, which aren't limited by a rigid target function, as they are generated and defined as the training process evolves. Another differentiator of this kind of model is that they can take a large variety of data types as input, such as speech, images, video, text, and other data susceptible to being represented as vectors.
AI is the last step in the scale of abstraction capabilities that,
in a way,
include all previous algorithm types, but with one key difference: AI algorithms are able to apply the learned knowledge to solve tasks that had never been considered during training. The types of data with which this algorithm works are even more generic than the types of data supported by machine learning, and they should be able, by definition, to transfer problem-solving capabilities from one data type to another, without a complete retraining of the model. In this way, we could develop an algorithm for object detection in black and white images and the model could abstract the knowledge to apply the model to color images.
In the following diagram, we represent these four stages of development towards real AI applications:
Types of machine learning
Let's try to dissect the different types of machine learning project, starting from the grade of previous knowledge from the point of view of the implementer. The project can be of the following types:
Supervised learning
: In this type of learning, we are given a sample set of real data, accompanied by the result the model should give us after applying it. In statistical terms, we have the outcome of all the training set experiments.
Unsupervised learning:
This type of learning provides only the sample data from the problem domain, but the task of grouping similar data and applying a category has no previous information from which it can be inferred.
Reinforcement learning:
This type of learning doesn't have a labeled sample set and has a different number of participating elements, which include an agent, an environment, and learning an optimum policy or set of steps, maximizing a goal-oriented approach by using rewards or penalties (the result of each attempt).
Take a look at the following diagram:
Grades of supervision
The learning process supports gradual steps in the realm of supervision:
Unsupervised Learning doesn't have previous knowledge of the class or value of any sample, it should infer it automatically.
Semi-Supervised Learning, needs a seed of known samples, and the model infers the remaining samples class or value from that seed.
Supervised Learning: This approach normally includes a set of known samples, called training set, another set used to validate the model's generalization, and a third one, called test set, which is used after the training process to have an independent number of samples outside of the training set, and warranty independence of testing.
In the following diagram, depicts the mentioned approaches:
