This document provides an overview of machine learning and how it can benefit businesses. It begins with defining machine learning as software that can learn from data like humans do in order to solve problems. The document then discusses myths and facts about machine learning, how it works, case studies of companies using it, and provides a guide for getting started with machine learning including adjusting mindsets, defining problems, collecting data, and finding tools. The overall message is that machine learning can provide competitive advantages and dramatically impact businesses if leveraged properly.
2. Samuel Adeshina
I have built modern platforms
and I know a thing or two about
Big Data.
Meet The Speaker
ML ENGINEER ( @ruralfarmershub )
3. So, Machine Learning?
It is a bit difficult to narrow down one specific
definition of machine learning (ML) because you’ll
get a different explanation depending on whom you
ask.
In its most primal sense, it is an attempt to invent
computer software that can think like we humans do.
What is unique about it though is the fact that it can
be leveraged as a solution to any kind of problem as
long as there is sufficient data for it.
4. Learning in itself is intimately human.
Humans Machines (Hardware and Software)
Biological Intelligence
Helpless infants learn from their
environment and those around them
and they improve over time.
Logical Reasoning
A set of step-by-step rules
that are designed for
machines to follow
Machine Learning
Machines are being taught to
learn about the world around
them just like the human
infants
Artificial Intelligence
The counterpart of an adult
human!
6. The goal of this presentation is
to tell you how
you can begin to leverage on
the power and potential of this
technology.
7. Focus Areas
Our discussion will be centred around the following:
➔ The Machine Learning Hype
Facts and Myths. You will learn what is possible and the
little that are not.
➔ Case Studies
You will learn how this technology is being harnessed
by the bold problem solvers of the world.
➔ How It Works
The Fundamentals. You will discover why data is a big
deal.
➔ A Getting Started Guide
You will discover how you can start using machine
learning as a tool for growth
8. Meet Jemila.
Jemila is a smallholder farmer that
grows and sells tomatoes and onions.
After every farming season, she sorts
her harvest and discover her yield by
separating the good harvests from the
bad ones and ensures that onions are
not mixed with tomatoes in storage.
She does this by physically inspecting
the products while using her years of
experience to tell tomatoes apart from
onions.
9. Despite Jemila’s ability to work very
quickly and the communal support she
often gets, there’s only so much she can
get done.
Therefore, her business is always on the
smaller-scale and barely has a footprint on
the agri forex market compared to
AgriCorp - a company running a business
that is similar to hers.
10. Meet AgriCorp.
AgriCorp owns expansive estates of
farmland across the country that
grows only tomatoes and onions.
Just like Jemila, AgriCorp needs to sort
their harvest too. Unlike Jemila, they
acquired an automated solution that is
able to tell the difference between a
tomato and an onion without human
involvement. This is integrated as part
of their harvest pipeline.
Story for illustration purposes only
11. AgriCorp’s Advantage
Their innovative solution is able to tell the difference between
tomatoes and onions by using its sensors to measure things like
the weights, surface roughness, moisture content, and so on of
the produce.
It was trained to recognize the combinations that represent an
onion and the ones that represents tomatoes. This was done by
feeding it lots of existing data to learn from.
The more harvest it sorts, the smarter it gets because it has
been exposed to more data.
12. AgriCorp’s solution is a machine learning approach that
does exactly what Jemila has to manually do.
Is this a disadvantage for Jemila?
13. Machine Learning used to take place behind the scenes
and was left alone for organizations like Amazon, Google
and Facebook to deal with.
But now machine learning is on the front pages of our
social timelines, is the subject of heated debates and is
shaping the future.
Unfortunately, all the hype have been accompanied by
lots of myths about its impacts.
14. Myths and Facts
Machine Learning
Can Be Used
Anywhere
Machine Learning
will take over
human work
Machine Learning
Platform Is Easy
To Build, And
Anyone Can Do It
MYTH MYTH FACT
15. Facts and Myths (Contd.)
Machine Learning
developments
represent ‘giant
leaps’ for
technology
Machine Learning
is autonomously
superior to
humans
Machine Learning
can do anything
with massive
amounts of data
MYTH MYTH FACT
16. Industries affected by Machine Learning
Seriously.
The amount of data in the world is
growing at an exponential pace.
It seems like machine learning is the
most efficient way to understand all of it.
There is literally an entire field dedicated
to the processing and extraction of
knowledge from data (conveniently
called data science.)
All of them, next.
17. Benefits of Incorporating ML
Outlined below is a list of different business processes Machine Learning
can have a positive impact on:
➔ Sales
➔ Customer Service
➔ Product Recommendations
➔ Automating Business Processes
➔ Improved Security
➔ Accounting
18. Three Crucial Areas
Creating Real
Business Value
Machine
Learning–Based
Automation
Efficient and
Real-Time
Decision Making
Where Machine Learning Can Benefit Small - Medium - Scale Businesses
19. Imagine if, in 5 years,
ML/AI technologies are
as integrated into
society as phones are
today
20. Machine Learning, How?
At its core, machine learning is about taking in input
data and processing it, learning to recognize patterns
and figure out what’s useful from that data to make a
helpful decision.
21. Supervised Learning
In supervised learning, we train an
agent using labeled data.
We can train a machine to process
and output a decision that matches
the expected output as accurately as
possible.
By feeding a machine inputs over
and over and having a penalty for
getting the wrong answer, we can
nudge it toward the correct answer.
Once the machine is properly trained,
it can give the right answer every
time to match our expected output.
22. Unsupervised Learning
In unsupervised learning, we have to
get machines to generate useful output
for problems we don’t already have the
answer to.
Machines try finding trends and
patterns when given unlabeled data as
input.
The process of recognizing patterns
and trends is how a learning algorithm
is able to make sense of the data.
Unsupervised learning is super
powerful in the real world.
24. Your starting
point is your data
The power of data in
solving everyday
problem is hinged on
the predictability of man
PEOPLE ARE MORE
PREDICTABLE THAN
PARTICLES
Data is your most important business
asset. You must first fully adapt to a
“good data beats opinion” philosophy to
harness Machine Learning.
25. Business owners must work with a
‘good data beats opinion’ philosophy.
As businesses grow, more and more people have opinions
about which steps to be taken. But it helps to utilize this
philosophy. Every metric of business can be tested,
measured and improved.
You need to have real-time access to the most important
data in your endeavours. For example, knowing which KPI’s
influence revenue and profit is much more important than
the revenue and profit themselves.
26. Amazing Local Case Studies
➔ Kudi.ai has developed a chatbot that uses AI to
understand user requests, drive conversations,
understand user spending habits and prevent fraud.
➔ Lara.ng has also developed a chat bot for public
transportation. It assists users get from one point to
another by providing detailed, text-based, step-by-step
directions and accurate fare estimates.
➔ Voyc assists quality assurance personnel by monitoring
contact centre interactions and provides valuable insights.
➔ Retina-AI, Drugger, and Vetsark
➔ WHAT ARE YOU WAITING FOR?
➔ Accounting
27. Getting Started
Adjust Mindset
- What is holding
you back?
- Find your ML Tribe
(1) (2) (3)
Pick A Process
- Define your
problem?
- What data does it
need?
Pick A Tool
- How do you
collect and
manage these
data?
28. Talk to an expert for
implementation.
OR
Become an expert.
You would have already been armed with the
prerequisites, the domain-centric knowledge and an
understanding of the solution you need.
29. Be Ready.
It’s a common fear that ML/AI will take away lots
of jobs, no profession or career will be spared.
But the great news is, an equal amount of jobs
will be created in the lateral direction.
Build A Porfolio
30. It seems plausible that in the near future AI will be
like electricity.
Every business in every industry will be using it
within the next 5 years or they’ll fall behind real fast.
It’s like how today you don’t ask corporations if they
use electricity or not; they just have to.
31. Competitive Advantage
Is the ability to learn more about our
customers faster than the competition and
the ability to turn that learning into action
faster than the competition.
32. The practical applications of machine learning drive results
which can dramatically affect a business’ bottom line.
Already signs are ripe that ML is going to play a major role
in offering a level playing field to the small and medium-
sized businesses of the future.
That said, it is important to avoid the temptation of running
for an ML-based solution just to get duped by the hype.
Instead, you should actually consider the specific ways
these solutions can add value to your (business) process.
33. Good luck!
On Your Machine Learning Journey. I hope
you will use this information to finally start
exploring the plethora of opportunities in your
domain.
For more information relevant to
this presentation, feel free to
send an email through the
contact form at:
bitly.com/samueladeshina
Editor's Notes
Automation helps businesses dealing with a scarcity of resources and manpower. This is more applicable to small enterprises that always operate with limited means and human capital. Machine Learning tools being embedded into software products can help businesses gather the most relevant insights to serve their customers appropriately. For instance, insurance companies can implement Machine Learning technology to suggest products to customers based on customer insights.
As the volumes and variety of user data continue to grow at a rapid speed, Machine Learning–based algorithms and tools now have better scope to gain customer insights by learning from this data. This in turn helps businesses make faster and often real-time decisions concerning their business operations and customer services. Machine Learning technology can easily facilitate connecting a business with streams of business data and thereby can help more precision-driven as well as fast-paced decision making.
Machine Learning technology continues to create business value for enterprises of all sizes and niches. In this respect, we need to understand how this technology can actually help customers and users with substantial business value that couldn’t be possible before the technology was adopted by businesses. Simply put, the pace of decision making, adhering to the business rules, and addressing customer needs with precision have all become possible thanks to sophisticated Machine Learning–based tools.