Enhancing Worker Digital Experience: A Hands-on Workshop for Partners
AI 101 for km washington november 2018 km world workshop
1. AI 101 for KM
KM World Workshop
Gordon Vala-Webb
November 2018
Washington, D.C.
Gordon@BuildingSmarterOrganizations.com
@BuildSmarterOrg www.BuildingSmarterOrganizations.com
2. What I think KM should be . . .
• Predictions & Bets
• Flows of info, ideas
decisions
• Designing / Nurturing
@BuildSmarterOrg www.BuildingSmarterOrganizations.com 2
3. Why does AI
matter?
Workforce
Healthcare
Fourth industrial
revolution
@BuildSmarterorg www.BuildingSmarterOrganizations.com 3
World Economic Forum
Mapping Global Transformation
https://www.weforum.org/about/transformation-maps
3
4. Annual value estimated $3.5 – $6.8 trillion
@BuildSmarterorg www.BuildingSmarterOrganizations.com 4
NOTES FROM THE AI FRONTIER INSIGHTS FROM HUNDREDS OF USE CASES McKinsey Global Institute
https://www.mckinsey.com/~/media/mckinsey/featured%20insights/artificial%20intelligence/notes%20from%20the
%20ai%20frontier%20applications%20and%20value%20of%20deep%20learning/mgi_notes-from-ai-
frontier_discussion-paper.ashx
5. Is AI magic?
@BuildSmarterorg www.BuildingSmarterOrganizations.com
“Any sufficiently advanced technology is indistinguishable from magic.”
Arthur C. Clarke
5
6. Is it a flying car?
@BuildSmarterorg www.BuildingSmarterOrganizations.com 6
By Eslivb [CC BY-SA 4.0 (https://creativecommons.org/licenses/by-
sa/4.0)] , from Wikimedia Commons
By Thesupermat - Own work, CC BY-SA 3.0,
https://commons.wikimedia.org/w/index.php?curid=23885520
Or is it just a car?
7. Where is Machine Learning on the Gartner Hype Cycle?
@BuildSmarterorg www.BuildingSmarterOrganizations.com 7
https://www.gartner.com/en/research/methodologies/gartner-hype-cycle
?
?
?
8. What is Machine Learning?
Creating algorithms
that can recognize patterns in large,
evolving data sets, and
drawing conclusions from past
experience by using that data
@BuildSmarterorg www.BuildingSmarterOrganizations.com 8
Machine Learning and Predictive Systems; World Economic Forum
https://toplink.weforum.org/knowledge/insight/a1Gb0000000pTDREA2/explore/dimension/a1Gb00000017L8jEAE/summary
10. Components of custom-built AI solution
•Technologies
•Tools /methodologies
•Business knowledge and experience
•AI expertise
•Focus on specific business opportunity
@BuildSmarterorg www.BuildingSmarterOrganizations.com 10
11. What is Machine Learning?
Creating algorithms
that can recognize patterns in large,
evolving data sets, and
drawing conclusions from past
experience by using that data
@BuildSmarterorg www.BuildingSmarterOrganizations.com 11
Machine Learning and Predictive Systems; World Economic Forum
https://toplink.weforum.org/knowledge/insight/a1Gb0000000pTDREA2/explore/dimension/a1Gb00000017L8jEAE/summary
12. IF this THEN that ELSE: Employee or Contactor?
@BuildSmarterorg www.BuildingSmarterOrganizations.com 12
Will the service be provided
from home?
Or will the services be provided in a determined
place (offices or facilities of the hiring company)?
Will the service be subject to a work shift or schedule?
If yes – is employee If yes – is employee
If no - is independent contractor If no:
Will the person be using uniform of the hiring
company?
If yes – is employee
If no:
Will the person be using the tools of the hiring
company?
If yes – is employee
14. What is Machine Learning?
Creating algorithms
that can recognize patterns in large,
evolving data sets, and
drawing conclusions from past
experience by using that data
@BuildSmarterorg www.BuildingSmarterOrganizations.com 14
Machine Learning and Predictive Systems; World Economic Forum
https://toplink.weforum.org/knowledge/insight/a1Gb0000000pTDREA2/explore/dimension/a1Gb00000017L8jEAE/summary
15. “Recognize patterns in data sets”
@BuildSmarterorg www.BuildingSmarterOrganizations.com 15
Principal Component Analysis (PCA)/SVD
Least Squares and Polynomial Fitting
Constrained Linear Regression
K-Means Clustering
https://dzone.com/articles/ten-machine-learning-algorithms-you-should-know-to
17. Training systems:
False positives
and negatives
A face is a face – isn’t it?
@BuildSmarterorg
www.BuildingSmarterOrganizations.com
17@BuildSmarterorg www.BuildingSmarterOrganizations.com
18. Is a roc a rock?
@BuildSmarterorg www.BuildingSmarterOrganizations.com 18
By Charles Maurice Detmold (1883-1908) - http://boards.collectors-
society.com/ubbthreads.php?ubb=showflat&Number=4741639, Public Domain,
https://commons.wikimedia.org/w/index.php?curid=20210405
26. Caveat – is AI fair?
•US sentencing guidelines – COMPASS – racially biased
•Siri giving inadequate instructions to women’s health
services
•Natural language models that associate ‘woman’ with
‘receptionist’
•US law enforcement crime prediction - PredPol –
racially biased
•Google image search for “CEO” – gender biased
@BuildSmarterorg www.BuildingSmarterOrganizations.com 26
27. Some KM use cases
• Customized / personalized search
• K (customer) support “bots”
• Text analytics and NLP
• Life sciences – diagnoses and finding cures
• Cybersecurity – attacks and effective responses
@BuildSmarterorg www.BuildingSmarterOrganizations.com 27
28. Customized / personalized search – e.g. Attivio
“A system . . . begins to
understand how Bob is great
at handling queries about a
problem with mobile while
Mary excels at database
issues . . . a support ticket
can be directed to the right
person in real time based . .
. on an intake call or email.”
@BuildSmarterorg www.BuildingSmarterOrganizations.com
28
https://www.attivio.com/blog/post/solving-customer-issues-
intelligent-answers-and-insights
29. @BuildSmarterorg www.BuildingSmarterOrganizations.com 29
10 Best Chatbots [services] of 2018 Consumers Advocate https://www.consumersadvocate.org/chatbots/a/best-chatbots
Knowledge support – e.g. Dialogflowhttps://cloud.google.com/blog/products/gcp/dialogflow-enterprise-edition-is-now-generally-available
https://medium.com/swlh/how-to-build-a-chatbot-with-dialog-
flow-chapter-1-introduction-ab880c3428b5
https://hackernoon.com/how-to-train-your-robot-ai-for-
everyone-69b96ad943e5
30. Build a bot
https://landbot.io/
@BuildSmarterorg www.BuildingSmarterOrganizations.com 30
More advanced!
Language Understanding (LUIS)
https://www.luis.ai/welcome
• Domain
• Intents
• Utterances
• Entities
32. Data is
everything!
• Structured
• Time series
• Text
• Audio
• Video
• Image
@BuildSmarterorg www.BuildingSmarterOrganizations.com
https://www.mckinsey.com/~/media/mckinsey/featured%20insights/artifi
cial%20intelligence/notes%20from%20the%20ai%20frontier%20applicatio
ns%20and%20value%20of%20deep%20learning/mgi_notes-from-ai-
frontier_discussion-paper.ashx
NOTES FROM THE AI FRONTIER
INSIGHTS FROM HUNDREDS OF USE
CASES McKinsey Global Institute
32
33. Some key challenges
•Massive data sets
• Labelling training data
• Large and comprehensive enough
• Organizational change
•Explaining results
•Focused – not generalizable learning
•Potential bias
@BuildSmarterorg www.BuildingSmarterOrganizations.com 33
NOTES FROM THE AI FRONTIER INSIGHTS FROM HUNDREDS OF USE CASES McKinsey Global Institute
https://www.mckinsey.com/~/media/mckinsey/featured%20insights/artificial%20intelligence/notes%20from%20the%20ai%20frontier%20applications%20and%20value%20of%20deep%20learning/mgi_notes-from-ai-
frontier_discussion-paper.ashx
36. VUCA-Digital world
cc: frogthroat - https://www.flickr.com/photos/22980078@N04
“The Specter of an Accidental China-U.S. War”
“Sears tanked because the company
failed to shift to digital.”
36
37. Our organizations optimized for “old” world
@BuildSmarterOrg www.BuildingSmarterOrganizations.com 37
Old world
Grow big
Reliably repetitive
Control
VUCA – Digital world
Grow adaptable
Radically responsive
Predict
39. Uncertainty / Ambiguity / Complexity =
What is happening?
@BuildSmarterOrg www.BuildingSmarterOrganizations.com 39
Classifying Experimental Designs
Source: https://www.socialresearchmethods.net/kb/expclass.php
40. “Try or try not. There is only do not with no try."
Gordon Vala-Webb – Building Smarter Organizations 2016 Slide 40
“Try or try not.
There is only
do not with no
try."
40
@BuildSmarterorg www.BuildingSmarterOrganizations.com