Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
SlideShare a Scribd company logo
www.beaufortfairmont.com 984.244.2313
Machine Learning
& How It Affects Testers
PAUL MERRILL
@dpaulmerrill
www.beaufortfairmont.com 984.244.2313
Agenda
What is Machine Learning?
How will it affect what we test?
How will it affect how we test?
Will it take our jobs?
www.beaufortfairmont.com 984.244.2313
I am not a Machine Learning expert!
www.beaufortfairmont.com 984.244.2313
What is Machine Learning?
www.beaufortfairmont.com 984.244.2313
Machine Learning
Machine learning is the subfield
of computer science that gives
computers the ability to learn
without being explicitly
programmed.
~ Arthur Samuel
(Paraphrased and attributed)
www.beaufortfairmont.com 984.244.2313
Machine Learning
A computer program is said to learn from
experience E with respect to some class of
tasks T and performance measure P if its
performance at tasks in T, as measured by P,
improves with experience E.
~ Tom Mitchell
www.beaufortfairmont.com 984.244.2313
3 Types of Learning
1.Supervised Learning
Inputs relate to outputs
Training
2.Unsupervised Learning
Anomalous data
Interesting relationships
3.Reinforcement Learning
Trial and error
Results as feedback
www.beaufortfairmont.com 984.244.2313
Which type of problems?
Natural Language Processing
Healthcare
Imaging
Transportation
Search
www.beaufortfairmont.com 984.244.2313
What are we using ML for?
Projections
Augmenting thought
Replacing or augmenting human judgement
Understanding data
Dealing with large amounts of data
Strategic advantage
www.beaufortfairmont.com 984.244.2313
How will ML affect what we test?
www.beaufortfairmont.com 984.244.2313
- World Quality Report 2016-17
www.beaufortfairmont.com 984.244.2313
What We Test Will Change
Runtime generated rulesets
Expectations less clear
Less repeatability
>0% error tolerance
No longer an exact answer
www.beaufortfairmont.com 984.244.2313
What We Test
Testing will get harder.
~ George Neal
Chief Data Analyst, PrecisionLender
www.beaufortfairmont.com 984.244.2313
The Future
For people who don’t like to do what
humans do well, the future is a very scary
place.
~ George Neal
www.beaufortfairmont.com 984.244.2313
How will ML affect how we test?
www.beaufortfairmont.com 984.244.2313
www.beaufortfairmont.com 984.244.2313
The Future
Machines are better
programmers than testers.
~ Jason Arbon, CEO, AppDiff
www.beaufortfairmont.com 984.244.2313
The Future
[You] have to be smarter than most
programmers to be a really good tester.
~ Jason Arbon
www.beaufortfairmont.com 984.244.2313
The Future
I still think the hardest thing in software is
testing!
~ Jason Arbon
www.beaufortfairmont.com 984.244.2313
The Future
When computers generate code, they can’t
test it either.
~ Jason Arbon
www.beaufortfairmont.com 984.244.2313
The Future
ML will augment our ability to be
extremely smart about what we
cover.
- Anonymous
www.beaufortfairmont.com 984.244.2313
The Future
[ML will work] within the
program, informing us of
what to test.
- Anonymous
www.beaufortfairmont.com 984.244.2313
The Future
There are a set of
problems that have
required human judgement
that are now tractable.
~ Doug Kubel
VP Product Development
UltraLinq Healthcare Solutions
www.beaufortfairmont.com 984.244.2313
The Future
[ML] may enable some types of testing.
~ Doug Kubel
www.beaufortfairmont.com 984.244.2313
Will ML take our jobs?
www.beaufortfairmont.com 984.244.2313
Testers vs. testers
10% vs. 90%
www.beaufortfairmont.com 984.244.2313
The Future
There is no way ML will be involved in every
single testing job in the next 30 years.
~ Anonymous
www.beaufortfairmont.com 984.244.2313
The Future
Deep learning won’t put everyone out of a
job,
but it will have an impact.
~ Doug Kubel
www.beaufortfairmont.com 984.244.2313
The Future
[ML] won’t put everyone out of jobs -
not in software development,
not in testing.
~ Doug Kubel
www.beaufortfairmont.com 984.244.2313
The Future
When asked, “Will ML replace testers?”
No.
~ Jason Arbon
www.beaufortfairmont.com 984.244.2313
I’ll leave you with one question…
www.beaufortfairmont.com 984.244.2313
The machines are learning,
are you?
www.beaufortfairmont.com 984.244.2313
References
“Machine Learning and AI Spending to Surge Toward $47B by 2020: IDC” -
Which-50, https://which-50.com/machine-learning-ai-spending-surge-toward-47-
billion-2020-idc/
“Some Studies in Machine Learning Using the Game of Checkers.” - Arthur Samuel,
http://citeseerx.ist.psu.edu/viewdoc/download?
doi=10.1.1.368.2254&rep=rep1&type=pdf
“Some Studies in Machine Learning Using the Game of Checkers. II-Recent
Progress ” - Arthur Samuel, http://researcher.watson.ibm.com/researcher/files/us-
beygel/samuel-checkers.pdf
“Machine Learning with Python” parts 1-57 - Harrison Kinsley, https://
www.youtube.com/watch?
v=OGxgnH8y2NM&list=PLQVvvaa0QuDfKTOs3Keq_kaG2P55YRn5v
www.beaufortfairmont.com 984.244.2313
References
“How to Use Machine Learning Algorithms in Weka” - Jason Brownlee,
http://machinelearningmastery.com/use-machine-learning-algorithms-weka/
“Weka 3 Data Mining Software in Java” - The University of Waikato, et al.,
http://www.cs.waikato.ac.nz/ml/weka/
“World Quality Report 2016-2017” - CapGemini, Hewlett Packard
Enterprise, Sogeti
“Machine Learning” - Wikipedia, https://en.wikipedia.org/wiki/
Machine_learning
“Machine Learning” - Tom M. Mitchell, McGraw-Hill, 1997
AppDiff - https://www.appdiff.com/
www.beaufortfairmont.com 984.244.2313
Special Thanks
George Neal, Chief Analytics Officer,
PrecisionLender
Jason Arbon, CEO, AppDiff
Doug Kubal, VP Product Development, UltraLinq
Healthcare Solutions
Anonymous, a Major Health Analytics Company
Anonymous, a Major Health Analytics Company
www.beaufortfairmont.com 984.244.2313
Call me today &
Talk through your challenges
984.244.2313
Paul Merrill
Webinars: beaufortfairmont.com/webinars
Podcast: Reflection As A Service
Email: paul@beaufortfairmont.com
Twitter: @dpaulmerrill

More Related Content

Quality Jam 2017: Paul Merrill "Machine Learning & How it Affects Testers"