Professional Documents
Culture Documents
Machine Learning and Blockchain
Machine Learning and Blockchain
Blockchain
The power of convergence
Jason Scott
Contents
Machine Learning Python and Blockchain
1. INTRODUCTION
2. WHAT IS MACHINE LEARNING?
3. WHAT IS ARTIFICIAL INTELLIGENCE?
4. WHAT ARE ARTIFICIAL
NEURAL NETWORKS?
5. THE RELEVANCE OF ARTIFICIAL INTELLIGENCE
6. ALGORITHMS AND BIG DATA AND HOW THEY CAN
INFLUENCE OUR CONSUMPTION CHOICES
7. ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING: A
WAY OUT
8. MACHINE LEARNING WITH PYTHON
9. THE FUTURE OF AI AND
MACHINE LEARNING
10. THE FUTURE OF MACHINE LEARNING IN VARIOUS
AREAS
11. WILL MACHINE LEARNING
COME TO AN END?
12. BLОСKСHАIN TЕСHNОLОGУ
13. WHY COMBINE BLOCKCHAIN AND MACHINE LEARNING
14. HOW THE AI CAN CHANGE BLOCKCHAIN
15. CONCLUSION
Thanks to Luìs Bigést, my friend and philosopher. Without his contribution during
the happy nerdy days, this work would not have existed
1. INTRODUCTION
The world today has become dynamic and transcending.
Everything as we know it is changing. Newer and newer editions of
versions of everything come up every day and this change is not
only limited to business, the workplace, education, industry, the
economy and the society at large and not to forget technology.
Technology has become a huge part of our lives and to most of us,
second nature. The relevance of technology today cannot be over-
emphasized. In the dispensation we live in now, computers and
machines in general are more than useful.
They have become integral in our daily lives, businesses,
education, agriculture, to name but a few. These days, a lot of things
can be done easier and faster. The weather, for instance can be
easily and near-accurately forecasted through the use of computers.
Advertising is another progressive field that technology has turned
around. The internet might just be the biggest of all the benefits of
technology.
Technology is going nowhere; it is here to stay. It has changed the
way we live, think, work, etc. Technology is taking the human race to
places even the human mind can hardly comprehend.
Computers and Machines are multi-purpose and they have
brought our society together into a global village and have changed
our lives.
One of the mind-blowing offspring of computer and computer
technology is Artificial Intelligence and Machine Learning.
Machine learning involves the study of algorithms and statistical
models that computer systems make use of to execute laid down or
programmed tasks without the use of explicit instruction.
Artificial intelligence involves the theory and development of
computer systems able to perform tasks that normally requires
human intelligence, such as visual perception, speech recognition,
decision-making, and translation between languages.
This book covers issues regarding machine learning and all you
need to know about machines and how they take and use
instructions. Be prepared for an amazing ride. Happy reading!
2. WHAT IS MACHINE LEARNING?
Machine learning is an application of artificial intelligence
(abbreviated and known popularly as AI) that provides systems the
ability to learn and improve automatically from its own history of
experiences without being explicitly programmed.
The title ‘Machine Learning’ was coined by Arthur Samuel in 1959.
A more formal definition of the algorithms studied in the field of
Machine Learning was put forward by Tom M. Mitchell. He quoted; ‘A
computer program is said to learn from experience E with respect to
some class of tasks T and performance P, improves with experience
E. Arthur Samuel, an American pioneer in the field of computer
gaming and artificial intelligence, coined the term while he was at
IBM.
A book on Machine Learning by Nilsson was made during the
1960s and discussed mostly about pattern classification. This
interest continued into the 1970s and was described in the book of
Duda and Hart in 1973. Machine learning stemmed out of artificial
intelligence which was a quest back then.
It was a scientific endeavor and was already an academic
discipline as some people were already interested in having machine
learn from data. They attempted to approach the problem with
different symbolic methods, as well as what were then termed
‘neural networks’, these were mostly perceptions and other models
that were later found to be reinventions of the generalized linear
models of statistics.
They also employed probabilistic reasoning, especially in
automated medical diagnosis. With time, there was an increasing
importance on the logical, knowledge-based approach and this
divided artificial intelligence and machine learning. The probabilistic
systems were plagued by theoretical and practical problems of data
acquisition and representation.
Machine learning today is a separate field from Artificial
intelligence and started to boom in the 1990s. The field changed its
goal from achieving artificial intelligence to tackling solvable
problems of a practical nature. Today, Machine Learning is relevant
in the fields of Agriculture, Anatomy, Adaptive websites, Banking,
Bioinformatics, Computer networks, Credit card fraud detection, DNA
sequence classification, Economics, Financial market analysis,
Marketing, Robot locomotion, Telecommunication, Speech
recognition, Time series forecasting, etc.
It lays emphasis on computer development programs that can
access data and use it to learn for themselves. The rate at which the
concept of Machine Learning is going today, it is speculated that it
will be associated with almost everything in time to come.
Machine learning involves the study of algorithms and statistical
models that computer systems make use of to execute laid down or
programmed tasks without the use of explicit instruction.
Machine Learning relies on inferences and patterns instead of
explicit instructions. Machine Learning is a branch of Artificial
Intelligence. The algorithms of Machine Learning create a
mathematical model that is usually based on sample data (this
sample data is known as ‘training data’). These models are used to
make decisions and to predict without being explicitly set to perform
the task.
Machine learning is strongly associated to statistics requiring
computation (otherwise known as computational statistics). The
process of learning starts with observation. Data, such as samples of
a population, instruction and direct experience, examples, etc., make
up what can be observed and from observation, good or better
decisions and predictions can be made.
The major reason for Machine Learning is so that computers can
learn on their own without direct input from humans. Machine
Learning allows and makes possible the analysis of a large volume
of data. This is carried out faster and more accurately.
There are quite a number of Machine Learning methods. These
methods include; supervised machine learning algorithms,
unsupervised learning algorithms, semi-supervised learning
algorithms and reinforcement machine learning algorithms.
scikit-learn
This is the project used by the Spotify team to generate the
recommended song lists for users. This library includes
classification, regression and clustering algorithms as well as
support vectors, logistic regression and a Bayesian classifier. It is
mainly used for data mining and data analysis tasks.
TensorFlow
TensorFlow, this project has been realized by Google and has been
implemented in many projects in a very short time. The library was
designed to be used in neural network libraries to perform numerical
computing based on data flow graphs. The Magenta team uses it in
its portal to generate Music/Art affiliations.
Tehano
Tehano, developed by researchers at the University of Montreal in
Canada is one of the most popular bookstores. It is used to help
developers evaluate mathematical expressions, in fact it has been
implemented in several compilers in order to speed up the
conversion of sources.
Pattern
Pattern, is a web mining module that can be used for Data Mining
and Network Analysis, but its strong point is the natural language
processing, has various tools such as part-of-speech taggers, n-
gram search, sentiment analysis and supports WordNet. With these
features it is possible to create automated bots for the moderation of
comments able to learn new words and meanings day by day.
Nilearn
Nilearn. This Python module is based on the already mentioned
scikit-learn and focuses on statistical learning, predictive models and
connectivity analysis. These features make it particularly suitable for
projects focused on NeuroImaging data.
9. THE FUTURE OF AI AND
MACHINE LEARNING
Programmers use machine learning as an application that allows
for artificial intelligence (AI) to learn automatically and improve itself
from experience, without being directly programmed. With Machine
Learning, AI's can simply access data and make complex predictions
and decisions based on these data. So, what exactly is in store for
artificial intelligence especially in the field of machine learning, in the
nearest future. Do what we see in Sci-Fi movies accurately predict
how things will turn out?
It is predicted that in the nearest future, AI and ML would be
integrated into every area of the human civilization, from the most
basic and simplest of rituals, like eating, to complex and far reaching
practices like warfare.
To achieve this, every company, both profit and non-profit oriented
alike would have to utilize AI and ML on a much larger scale and for
cheaper cost. Cost of machine learning systems would be lowered,
while its capacity would be increased by the cloud service providers.
They have even built the Tensor Processing Unit (TPU), which
allows users to train their various machine learning software as
quickly and as easily as possible.
Today, we already see how machine learning influences us, from
social media and web browsers to research and data analysis tech.
but we still have a very long way to go before we can truly see the
effects of machine learning in AI as show cased in those Sci-Fi
movies.
While machine learning is a very good tool in improving artificial
intelligence, users have complained of its slow paced development;
the seemingly less complex task of identifying or predicting objects
and their names can be daunting as thousands of images of
examples of said object has to be shown.
10. THE FUTURE OF MACHINE
LEARNING IN VARIOUS AREAS
In the past, uncovering deep insights from machine learning and
artificial intelligence (AI) required extensive programming skills and
an intimate knowledge of math – experience in short supply in most
enterprises. A new class of machine learning tools, however, is
making it possible for existing business analysts and business
professionals, to take machine learning and AI initiatives into their
own hands by democratizing data science.
Machine Learning technology can protect the companies that are
dealing with finance, from financial fraud that may occur in the future.
Apart from this, machine learning can help to predict the upcoming
opportunities that could be implemented for further investments.
In the United States alone, over 40% of their companies make use
of machine learning in improving their sales and market. At least
75% of the companies mentioned above have exceeded their sales
target as a result of machine learning. Even banks in the European
economy have reduced their churn rates by 20% and managed to
increase sales by 10%, which says a lot in the financial sector.
This is as a result of machine learning. IDC research has shown
that 3 out of every 4 enterprises and developers would make use of
Machine learning in this year alone. Also, research has shown that a
minimum of 40% of digital improvements would be made worldwide
because of the use of machine learning.
From all these points, it is safe to conclude that Machine learning
will not be coming to an end both now and in the future, as far as the
industrial and digital world is concerned. It would play a major role in
supporting businesses of all size, hence, it would be for everyone
wise to invest in this cutting edge technology.
12. BLОСKСHАIN TЕСHNОLОGУ
A blосkсhаin iѕ a рubliс lеdgеr оf infоrmаtiоn collected thrоugh a
nеtwоrk that ѕitѕ on tор оf thе internet. It is how this infоrmаtiоn is
recorded that givеѕ blосkсhаin itѕ grоundbrеаking роtеntiаl.
Blockchain technology iѕ not a company, nоr iѕ it аn app, but
rаthеr аn еntirеlу new wау оf dосumеnting dаtа оn thе intеrnеt. Thе
tесhnоlоgу саn be used to dеvеlор blосkсhаin applications, such аѕ
ѕосiаl nеtwоrkѕ, mеѕѕеngеrѕ, games, exchanges, ѕtоrаgе рlаtfоrmѕ,
vоting ѕуѕtеmѕ, рrеdiсtiоn markets, оnlinе shops аnd much more. In
this ѕеnѕе, it iѕ similar tо thе intеrnеt, whiсh iѕ whу ѕоmе hаvе
dubbеd it “Thе Intеrnеt 3.0”.
Thе infоrmаtiоn recorded оn a blockchain саn tаkе оn any form,
whеthеr it be dеnоting a trаnѕfеr оf mоnеу, оwnеrѕhiр, a trаnѕасtiоn,
ѕоmеоnе'ѕ idеntitу, an agreement bеtwееn twо раrtiеѕ, or еvеn hоw
much еlесtriсitу a lightbulb has uѕеd. Hоwеvеr, to dо ѕо requires a
соnfirmаtiоn from ѕеvеrаl оf dеviсеѕ, ѕuсh аѕ computers, оn the
nеtwоrk. Onсе аn аgrееmеnt, оthеrwiѕе known аѕ a consensus, iѕ
rеасhеd bеtwееn thеѕе dеviсеѕ to ѕtоrе ѕоmеthing оn a blockchain
it iѕ unquestionably thеrе, it саnnоt bе diѕрutеd, removed or аltеrеd,
withоut thе knоwlеdgе and permission оf those whо mаdе that
rесоrd, аѕ wеll as thе widеr соmmunitу.
Blосkсhаin owes itѕ name tо hоw it wоrkѕ аnd the mаnnеr in
whiсh it ѕtоrеѕ data, nаmеlу that the information iѕ packaged intо
blocks, whiсh link to fоrm a сhаin with other blocks оf similar
infоrmаtiоn.
It iѕ thiѕ асt оf linking blосkѕ intо a сhаin thаt mаkеѕ thе
information ѕtоrеd оn a blосkсhаin so trustworthy. Onсе thе dаtа iѕ
recorded in a block it саnnоt bе altered withоut having tо сhаngе
every block that came аftеr it, mаking it impossible tо do ѕо withоut it
being seen by the other раrtiсiраntѕ оn thе nеtwоrk.
Blосkсhаinѕ are tаmреr evident and tаmреr rеѕiѕtаnt digitаl
lеdgеrѕ implemented in a distributed fаѕhiоn (i.e., withоut a сеntrаl
repository) аnd usually withоut a сеntrаl authority (i.e., a bаnk,
company, оr gоvеrnmеnt). At thеir basic lеvеl, thеу enable a
соmmunitу оf uѕеrѕ tо record transactions in a shared ledger within
that community, ѕuсh that undеr nоrmаl ореrаtiоn of thе blосkсhаin
network nо trаnѕасtiоn саn bе сhаngеd оnсе published. In 2008, the
blосkсhаin idеа wаѕ соmbinеd with ѕеvеrаl оthеr tесhnоlоgiеѕ and
computing соnсерtѕ to сrеаtе mоdеrn сrурtосurrеnсiеѕ: electronic
саѕh рrоtесtеd through сrурtоgrарhiс mechanisms instead of a
сеntrаl rероѕitоrу оr authority. The firѕt such blockchain bаѕеd
cryptocurrency was Bitсоin.
Within the Bitcoin blockchain, information representing еlесtrоniс
саѕh is attached to a digital address. Bitсоin uѕеrѕ саn digitally sign
аnd trаnѕfеr rightѕ tо thаt infоrmаtiоn to аnоthеr uѕеr аnd thе Bitсоin
blосkсhаin rесоrdѕ thiѕ transfer рubliсlу, аllоwing аll раrtiсiраntѕ оf
thе nеtwоrk tо indереndеntlу verify thе validity of thе trаnѕасtiоnѕ.
Thе Bitсоin blockchain iѕ ѕtоrеd, mаintаinеd, аnd соllаbоrаtivеlу
managed by a distributed grоuр оf раrtiсiраntѕ. This, аlоng with
сеrtаin сrурtоgrарhiс mechanisms, mаkеѕ thе blockchain resilient tо
attempts tо аltеr thе ledger lаtеr (mоdifуing blосkѕ оr fоrging
trаnѕасtiоnѕ).
Thеrе iѕ hype аrоund thе uѕе of blосkсhаin tесhnоlоgу, yet the
technology is nоt well undеrѕtооd. It iѕ nоt magical; it will nоt solve
all рrоblеmѕ. Aѕ with аll new technology, thеrе iѕ a tеndеnсу tо want
tо аррlу it to every ѕесtоr in every way imaginable.
Building trust
Cost reduction
Acceleration of transactions
Manufacturing
As part of the manufacturing process, companies have begun to
rely on bitcoin blockchain-based processes and intelligent contracts
to enable production controls, transparency, security and
compliance. Instead of planning fixed machine maintenance
programs, predictive algorithms, which are at the heart of machine
learning technology, are used to create flexible intervention plans,
right at the precise moments when they should occur. Quality control
and product testing have also become progressively automated, with
adaptive and computer vision algorithms successfully used to detect
good and defective products, especially in highly sensitive
environments.
[1] You can see “Beyond Bitcoin : Why Blockchain will change our Life, Economy and
the World” http://bit.ly/BlockchainRevolution2020