Documente Academic
Documente Profesional
Documente Cultură
01
WHO IS THIS
PROGRAM FOR
Participants will gain a practical understanding of the tools and techniques used in machine learning
applications. In the MIT tradition, you will learn by doing. There are no prerequisites in terms of math
or computational science, although basic understanding of statistics is helpful. This is not a coding
course, but rather an introduction to the many ways that machine learning tools and techniques can
help make better decisions in a variety of situations.
PARTICIPANT
TESTIMONIALS
02
YOUR
LEARNING JOURNEY
The program takes you through a progression of the four building blocks of machine learning. In
addition to the practical knowledge that you’ll gain from the video lectures, you’ll learn from live
webinars with faculty, peer discussions, case studies, assignments, and assessments. There are
application exercises—technical and non-technical; participants can choose exercises that work best
with their abilities and goals. Throughout the journey, there are real-life machine learning facilitators
to provide guidance and clarification. At the end of the program, you should be prepared to apply
some of these techniques at your own organization.
03
MACHINE LEARNING:
BUILDING BLOCKS
10110
010110
0100
Understand data
Data
1
Make predictions
Regression
2 Classification
Neural networks
Make decisions
under uncertainty
3
Foundations
Applications
4
Decisions
Determine causal
$ inference
Note: MIT Professional Education reserves the right to change the curriculum and delivery of the program to
improve the participant experience.
04
WHAT
YOU’LL LEARN
The tools and techniques in this machine learning program can help to address many common
challenges. Learn with examples from:
BANKING RETAIL
How do you predict To optimize your
whether a borrower inventory, how do
will default on a loan? you know whether
to pull from a
distribution center or
a retail store to fulfill
an online order?
PHARMACEUTICAL FINANCE
When developing How are data
new drugs, how can scientists exploring
you design better ways to predict
experiments to know the future price of
if a new drug will be digital assets, such
more effective than as Bitcoin?
an existing one?
MARKETING ECOMMERCE
How do you know How do you decide
which marketing when to cross-market
channel is performing vs. upsell a customer
best, and what is the at checkout—what
interaction effect drives more revenue?
when you are using
multiple channels?
05
PROGRAM
MODULES
Overview of the four building blocks of Learn the basic characteristics of data sets
machine learning. and identify effective statistical tools and
visualizations to glean insights from your data.
Understanding data: What is it telling us?
Ask the right questions of the data
Prediction: What will happen?
Know which tools to use to unlock insights
Decision making: What to do?
How data visualization clarifies data
Causal inference: Did it work?
Application: Wikipedia and Online
Marketplaces, Amazon, and Etsy.
Understand the basic concept of linear Classification is used to predict outcomes that
regression and how it can be used with fall into two or more categories, such as: male/
historical data to build models that can female, yes/no, or red/blue/green.
predict future outcomes.
Compare the ability of different methods to
How to build a model that fits best with minimize prediction errors
your data
Make better predictions, based on your data
How to quantify the degree of your uncertainty and desired outcome
What to do when you don’t have enough data Use the right approaches to deal with data
complexity
What lies beyond linear regression
Application: Using the data set provided, predict
Application: Using a provided data set, you'll the default risk of a pool of applicants in the credit
work through a scenario and answer application card industry.
questions.
06
Module 5: Prediction Part 3 – Neural Module 6: Decision Making
Networks Foundations
Neural networks are much like the networks Decision making is about selecting the
in the human brain. They are used in machine “optimal” decision or action in the presence
learning to model complex relationships of uncertainty, which all professionals face
between inputs and outputs and to find regularly.
patterns in data.
Learn how the decisions you make impact the
What are neural networks and how do immediate future and beyond
they work?
Choose the right approach based on the
Explore the history and examples of simple environment, rate of information flow,
and complex neural networks and your goal
How neural networks minimize errors, Strike the balance between exploration
regardless of the size of your data set (identifying what we don’t yet know) with
exploitation (using what we already know)
Application: Working in the Tensorflow sand-
box, you'll complete an exercise on neural Application: Retail planning, website design,
network architecture. and creating a perfect chess player.
Focus on many of the practical applications Causal inference is about understanding the
of decision making, from static to dynamic relationship between cause and effect.
situations.
Given a set of observations, identify what
Learn about making real-time caused those observations
recommendations for customers to drive
more business Learn to design experiments that deliver
meaningful conclusions
Learn how to achieve optimal inventory
management Application: Assess whether discount coupons
caused higher sales of certain products, or
See how data scientists are exploring ways whether some external factor, such as an
to predict the future price of digital assets improved economy, was the reason for the
such as Bitcoin increased sales.
07
PROGRAM
FACULTY
Devavrat Shah is a professor with the
department of electrical engineering and
computer science at MIT. He is a member
of the Laboratory for Information and
Decision Systems (LIDS) and Operations
Research Center (ORC), and the Director
of the newly formed Statistics and Data
Center in Institute for Data, Systems, and
Society (IDSS). His research focus is on
the theory of large complex networks,
which includes network algorithms,
stochastic networks, network information
theory, and large-scale statistical inference.
08
CERTIFICATE
Get recognized! Upon successful completion of the program, MIT Professional Education grants a
certificate of completion to participants. This program is scored as a pass or no-pass; participants
must receive 80 percent to pass and obtain the certificate of completion.
Participants will be awarded 6.4 continuing education units towards professional development.
E
Massachusetts Institute of Technology
L
This is to certify that
P
Your Name
Machine Learning
From Data to Decisions
(6.4 Continuing Education Units Awarded)
09
SPECIAL
ACHIEVEMENT AWARD
MIT Professional Education Fire Hydrant Award
It’s said that studying with MIT is like drinking from a fire hose—intense and satisfying. For those
participants who display extraordinary efforts in exhibiting that MIT ethos and demonstrate leadership
by going above and beyond in their program participation, they’ll be considered to receive the coveted
MIT Professional Education Fire Hydrant Award. This award can be displayed in professional bios,
such as on LinkedIn. Decisions are made by MIT faculty and the program team based on active
participation and behaviors that exemplify leadership and contribution to the overall program
experience for the cohort.
L E
This is to award a certificate of recognition to
M P
S A For all your extraordinary efforts in exhibiting the MIT ethos of
"drinking from the fire hose" and program participation in
Machine Learning
From Data to Decisions
10
DURATION 8 weeks, online
excluding orientation
6-8 hours/week
CONNECT WITH A
PROGRAM ADVISOR
Email: mit@emeritus.org
Phone: +1 315-982-5094
+1 315-277-2746