Connected Intelligence
Abel Sanchez & John Williams
MIT Campus | Cambridge, MA
University of Bologna - 1088
World Wide Web
Connecting Physical to Virtual
TRANSFORMING INDUSTRY
Field Intelligence
Finding the
Open
Receiver: A
Quantitative
Geospatial
Analysis of
Quarterback
Decision
Making
Jeremy Hochstedler
Spatial Reasoning
Police complaints drop 93 percent after
deploying body cameras
50Billion connected devices by 2020
Web 4.0
Shortcomings of the web
• Over 2 billion unplugged
• A large part of the world does not have access to the benefits
• Dominated by large corporations
• Products for those with access
• Privacy from corporations and government
• We are our data yet we do not control our data
• Censorship
• the great firewall of China
• Provenance
• Where does our food come from?
• Transparency and veracity
• Government accountability
• Fake news
Advantages
• No third parties
• No fees (lower fees)
• Faster
• Privacy
• Less Censorship
• More robust systems
• Foundation for services and products
Education’s Digital Transformation
Abel Sanchez & John Williams
MIT Campus | Cambridge, MA
Students &
Instructors
Master
Courses
Content
Delivery
Courses
Collaboration
Progress
Tracking
Assignments
Practices &
Exams
Hosting
Course
Navigation
Self
Registration
SCORM &
Standards
Learning
Management
System
Learning Management Systems
Learning Management Systems
Faculty Students
Learning
Management
System
Internal Publishing
Open Courseware (OCW)
Faculty
Content
Repository
External Publishing
No assessment, guidance, or credit
Massive Open Online Course
(MOOC)
Faculty Students
Learning
Management
System
Some assessment, guidance, no credit
Small Private Online Course
(SPOC)
Faculty Students
Learning
Management
System
Possible assessment, guidance, and credit
Active Learning
Other Models
MIT.x MicroMaster
Historical Admissions Model
Admission
Exams
Grades
EducationIntelligenceGrowth
Vision
Initiative
Drive
Character
Persistence
Community
Contributions
Creativity
Innovation
Leadership
Potential
Activities
Interests
Few Get In
Historical Admissions Model
MIT
Admission
MIT.x MicroMaster
Assessment, guidance, credit if admitted
High
Performers
no
yes
MIT
Admission
MIT Online
Course Work
Open to
Everyone
Open access.
Performance on MIT course work
determines success
Course Marketplace: Udemy
Nano degrees: Udacity
Bootcamps
DATA DRIVEN EDUCATION
We need to be faster
• We need to shorten the feedback cycle for students
• We need to shorten the content creation cycle for
faculty
• We need to have comprehensive analytics
360 Student View
Predictive Modeling for Student Interventions
Predict Performance On First Day
Help lower performing students
Challenge High Achievers
Explore Paths to Graduation
Student Wellness
EDX - BIG DATA
Speaking Speed
120 words per minute?
150 words per minute?
180 words per minute?
Speaking Length
60 minutes?
30 minutes?
15 minutes?
Students
preferred hand-
diagrams by a
“substantial-
margin”
MY CLASSROOM
Requirements Before Class
• Watch video
• Take mini-quiz
In Class
• 3 cycles of
• Mini-Lectures
• Active Learning
• Source code
• Tested
• Analyzed
• Best solutions identified
• Everyone walks out with a grade
Quizzes
• Authentic Examinations
• Students write/run programs
• Students
• Checkout code
• Submit solutions
• Are evaluated real-time
Forums
• Class policy, all questions must be posted on the
class forum (piazza).
• Solutions posted on forum
• Announcements posted on forum
• Participation pointes given for answering a peer’s
question
Data Collected
• Every keystroke:
• Video watching
• Pre-class quizzes
• Active learning code
• Homework
• Quizzes
• Forum activity
THE EDUCATION LANDSCAPE IS
CHANGING FAST
Video Content Explosion – Video Learning
Twitter
YouTube
Google Search
Google Docs/Drive
PowerPoint
DropboxFacebook
WordPress Skype
Evernote
Prezi
Wikipedia
Pinterest
LinkedIn
Moodle
iPad and Apps
Kahoot
Blogger
PowToon
Slideshare
WhatsApp
Google Chrome & Apps
Google Hangouts
Snagit
Audacity
Articulate Storyline
Screencast-O-matic
Yammer
Padlet
Word
Camtasia
Socrative
Khan Academy
Adobe Connect
TED Talks/Ed
Feedly
Canvas
Adobe Captivate
Edmodo
Google+
iSpring Suite
Diigo
Google Scholar
Coursera
SharePoint
OneNote
Explain Everything
Videoscribe
Pocket Nearpod
Office Mix
Gmail
Udutu
Google Translate
KeynoteExcel
JingAdobe Photoshop
Google Apps
Scoopit
Schoology
Outlook
GoAnimate
SurveyMonkey
Kindle & App
Google Maps
Notability
Google Sites
Quizlet
Sway
Vimeo
WebEx
Instagram
Firefox & Add-ons
iTunes and iTunesU
iMovie
Blackboard Collaborate
Movie Maker
Poll Everywhere
Tweetdeck
Canva
TrellSlacko
IFTTT
EDpuzzle
Flipboard
Udemy
Todays Meet
ThingLink
Easygenerator
Lectora Inspire
Haiku Deck
Piktochart
Adobe Acrobat DC
Blackboard Learn
Wordle
Mentimeter
SoftChalk
edX
DeliciousPiazza
100 Tools for Learning
Education Technology Players
Investment is Growing
Only 20% of today’s workforce
have the skills they’ll need for 60%
of the jobs that will exist in the
next five to 10 years.
[https://hbr.org/2016/09/the-solution-to-the-skills-gap-could-already-be-inside-your-company]
• Abel Sanchez
Connected Intelligence
Abel Sanchez & John Williams
MIT Campus | Cambridge, MA
How most companies build software is
broken, programmers write code in
single player mode
[https://techcrunch.com/2017/10/06/sourcegraph-raises-20m-bring-more-live-collaboration-to-programming/]
The Problem of Discovery
• Discovering, exploring, selecting, and consuming
are costly.
• The sole question of discovery can be daunting on a
single academic department.
• Extending the discovery to a campus, an
organization is a huge challenge.
Reproducibility
• 100 psychology experiments repeated, less than
half successful (Science 08/28/2015)
• 67 Economics papers 50% could no be
independently replicated (Federal Reserve 9/4/2015)
• 53 'landmark' cancer publications, majority cannot
be replicated (Nature, 3/28/2012)
Measuring Reproducibility in
Computer Systems Research
Current projects take too long,
especially data science projects
• In the research space, we continuously reinvent the
wheel.
Today’s Big Difference:
An Industry Built on Reuse
• Uber, Air Bnb, Amazon, etc
• An ecosystem of technologies to support building
fast, a new generation of technologies
A New Generation of Technology
• Cloud computation
• Agile Application Servers
• VMs – e.g. AWS
• Representational State
Transfer – REST
• Microservices
• Containers and Data
Stores
• Containers
• Package Management
• Data Stores
• Data Services
• Blockchain & Machine
Learning
• GIT, CI, CD
• DApps
• Merkle Trees
• Internet ML Services
• Hardware & Devices
• Commodity Hardware
• Async Computation
• Pub/Sub Streaming
Concepts
• Channels: data from a service
• Triggers (this): "trigger” based on condition
• Action (that): action taken on trigger input
• Recipes: are the predicates made from Triggers and
Actions
• Ingredients: are basic data made available from a
trigger
Connected Intelligence Platform
Our goal is analytics
• Faster feedback loops
• Identify the stars
• Accelerate knowledge creation
• Target investment
Q&A

Dr Abel Sanchez at Bristlecone Pulse 2017 MIT