Sps ottawa 2016 vincent biret - Microsoft graph and machine learning
-
Upload
vincent-biret -
Category
Technology
-
view
202 -
download
2
Transcript of Sps ottawa 2016 vincent biret - Microsoft graph and machine learning
![Page 1: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/1.jpg)
WELCOMESHAREPOINT SATURDAY
OTTAWA
December 3rd, 2016
![Page 2: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/2.jpg)
Vincent Biret
Make Graph Data useful for your company
![Page 3: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/3.jpg)
House keeping
![Page 4: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/4.jpg)
SPS Ottawa is made possible by our Sponsors!Platinum
Gold
Silver
Bronze
Bronze
![Page 5: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/5.jpg)
6 |SharePoint Saturday Atlanta
ShareP ntSummerhays Grill
5:30 pm1971 Baseline Road (corner of Woodroffe)
Please drink responsibly . We will be happy to call a cab for you
![Page 6: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/6.jpg)
Vincent BIRETOffice Servers And Services [email protected]/vince365
Products Team Tech Lead
Montreal
About Me
![Page 7: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/7.jpg)
8 |SharePoint Saturday Atlanta
Graph and Machine learning are going to be game changers for businesses in next 10 years
IOT is the next big wave
Not caring now would be like not caring about the cloud back in 2008
Why should you care?
![Page 8: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/8.jpg)
9 |SharePoint Saturday Atlanta
Users who are tired of “stupid” and isolated applications
Developers who want to ship awesome apps!
Deciders who want to make something out of their data
Who’s that session for?
![Page 9: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/9.jpg)
10 |SharePoint Saturday Atlanta
Understand what’s a/the graph Understand what are MS Graph and Delve Understand why it’s a game changer for your
business Learn how to use it in your applications Understand what’s Azure Machine learning Learn how to use it in your applications
Today’s objectives
![Page 10: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/10.jpg)
11 |SharePoint Saturday Atlanta
Graph Theory MS Graph Delve MS Graph API Machine learning theory MS Azure ML Conclusion
Agenda
![Page 11: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/11.jpg)
12 |SharePoint Saturday Atlanta
Ready?
![Page 12: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/12.jpg)
What is The Graph?
Graph Theory
![Page 13: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/13.jpg)
14 |SharePoint Saturday Atlanta
Is That a graph?
Category 1 Category 2 Category 3 Category 40
1
2
3
4
5
6
Title
Series 1 Series 2 Series 3
Sales
1st Qtr 2nd Qtr 3rd Qtr 4th Qtr
![Page 14: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/14.jpg)
15 |SharePoint Saturday Atlanta
That’s a Graph!
![Page 15: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/15.jpg)
16 |SharePoint Saturday Atlanta
RDBMS’s Suck!....
At doing what they are not meant for.
Why Graphs?
![Page 16: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/16.jpg)
17 |SharePoint Saturday Atlanta
The Property GraphVincent
Desk: E43
Phone: 514 444 4444
Extension: 275
Negotium
Street Address: Montreal
Creation : 1/1/00
Technical Advisor
Must do: technical advising
Advantages: better business cards
Developper
Must do: development
Advantages: better keyboard
Works asSince 1/7/14
Works asSince 12/7/12
![Page 17: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/17.jpg)
18 |SharePoint Saturday Atlanta
Graphs can be represented by matricesVery easy to compute by CPU’sLow memory usage
Why are computers so good with Graphs?
![Page 18: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/18.jpg)
The Microsoft Graph
![Page 19: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/19.jpg)
20 |SharePoint Saturday Atlanta
Data is in silosAccessing different workloads is hard
Search doesn’t workPoints out new things
Why a Microsoft Graph?
![Page 20: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/20.jpg)
21 |SharePoint Saturday Atlanta
What’s Microsoft’s Graph?
![Page 21: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/21.jpg)
22 |SharePoint Saturday Atlanta
WebHooksOpenType extensions
SharePoint (Sites/Lists/ListItems)
Org contactsDirectory
(everything in AAD)
Identity Protection
Tasks (planner)OneNote
Latest News?
![Page 22: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/22.jpg)
23 |SharePoint Saturday Atlanta
Graph.microsoft.ioResources
![Page 23: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/23.jpg)
Delve
Demo
![Page 24: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/24.jpg)
MS Graph API
![Page 25: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/25.jpg)
26 |SharePoint Saturday Atlanta
Data Nodes Actors Edges
Some Edges Modified Viewed TrendingAround WorkingWith OrgManager OrgColleague
Edges properties ActorId ObjectId Action Type Time Weight
Node properties SharePoint Search Schema Object model
Structure
![Page 26: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/26.jpg)
MS Graph API
Demo
![Page 27: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/27.jpg)
Machine Learning theory
![Page 28: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/28.jpg)
31 |SharePoint Saturday Atlanta
State of the art
![Page 29: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/29.jpg)
32 |SharePoint Saturday Atlanta
Machines can be trained to “guess stuff” “They” can get better at doing itNot AI but a step towards itNot that new to the business world
Highlights
![Page 30: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/30.jpg)
33 |SharePoint Saturday Atlanta
You have training data with expected results
You have control data with expected results
Build the experiment with a feedback loop
Train it
Put it in prod
Supervised learning
![Page 31: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/31.jpg)
34 |SharePoint Saturday Atlanta
Used to predict outcomes with few possible values
Eg “married”, “divorced”….Eg “rev > 50K”, “rev < 50k”…
Classification
![Page 32: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/32.jpg)
35 |SharePoint Saturday Atlanta
Used to predict continuous values
Eg Potential profit of somethingEg Potential time to achieve something
Regression
![Page 33: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/33.jpg)
36 |SharePoint Saturday Atlanta
You have data without expected results
Build the experiment with a feedback loop
Train it
Put it in prod
Unsupervised learning
![Page 34: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/34.jpg)
37 |SharePoint Saturday Atlanta
Used to detect natural grouping patterns of data(ie: data that might be related together)
Produces groups of data and puts the data in it
Clustering
![Page 35: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/35.jpg)
38 |SharePoint Saturday Atlanta
Ideal to match data together
Things likeMovies you might like Items others boughtOnline dating (matching you with another person)
…
« Matchmaker »
![Page 36: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/36.jpg)
With great power comes great responsabilities
Azure Machine Learning
![Page 37: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/37.jpg)
40 |SharePoint Saturday Atlanta
Now your applications can become “clever” !!!
Why so important to dev’s?
![Page 38: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/38.jpg)
41 |SharePoint Saturday Atlanta
Machine Learning* as a service
* Mostly predictive and semantic analytics
ML Studio
Not an Expert System
Highlights
![Page 39: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/39.jpg)
42 |SharePoint Saturday Atlanta
Get dataMake an experimentTest itGenerate a modelPublish an API Integrate with your App
Methodology
![Page 40: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/40.jpg)
ML Studio
Demo
![Page 41: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/41.jpg)
Time to day goodbye
Conclusion
![Page 42: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/42.jpg)
45 |SharePoint Saturday Atlanta
Better integration between apps/workloads (Graph)
Better understanding of the data by apps (and predictive) (ML)
Better user experience/productivity
Happier users
Money saved for the company
Conclusion
![Page 43: Sps ottawa 2016 vincent biret - Microsoft graph and machine learning](https://reader036.fdocuments.net/reader036/viewer/2022062311/589994b01a28ab30328b6589/html5/thumbnails/43.jpg)
46 |SharePoint Saturday Atlanta