Sldies I proposed for the lecture at CUOA Business School for the 2016 Edition of the Executive Master on ITC - Path Big Data e Social Analytics - http://www.cuoa.it/ita/formazione/corsi-executive/jobleader-big-data-e-social-analytics.php#/
1. @pieroleo
Transforming Data into Wisdom
Pietro Leo
Executive Architect - IBM Italy CTO for Big Data
Analytics & Watson
IBM Academy of Technology Leadership
Head of IBM Italy Center of Advanced Studies
2. 2
Tech Age
You$shared$
your$position$
with$me$and$
can$guess$
your$mobility$
need.$
I$can$take$you$
where$you$
need$to$be
Just$enjoy$
your$new$
experience
.$Stay$safe$
as$in$your$
home
I$know$
what$is$
needed$for$
you,$even$
before$you$
order$it
Please,$come$
with$me$and$
stay$by$me.
I$know$your$
content$I$can$
take$care$of$all$
your$digital$life
Has DATA'a'gravity?
Data'growth and'gravity distorts and'impacts
every component'of'IT'– and'business
Data & Big Data
Toward a Precise Decision Making to reduce the wasteful
spend as well as the risk in every industry
New Information Technology challenge is now about the
possibility to expand our WISDOM options
Watson
Wisdom
Ecosystem
and,Partners
Industry,
Solutions
Client
Solutions
&,products
IBM$
Provided
Data Publically
Sourced
Data
Partner
Provided
Data
Private
Client
Data
IBM,Watson,Innovation,platform,for,Cognitive,Business
Watson'Health
Watson'Financial'Service
Watson'Internet'of'Things
Hybrid,Watson,
Frameworks
Watson
Services,B API
Data
Knowledge
Wisdom
Cognitive Platform Cognitive Solutions
Anaphoric* Co,referencing
Colloquialism* Processing
Content* Management* ,, Versioning
Convolutional* Neural* Networks
Curation
Deep* Learning
Dialog* Framing
Ellipses
Embedded* Table* Processing
Ensembles* and* Fusion
Entity* Resolution
Factoid* Answering
Feature* Engineering
Feature* Normalization
Focus* and* Spurious* Phrase*
Resolution
HTML*Page* Analysis
Image* Management
Information* Retrieval
Knowledge* (Property)* Graphs
Knowledge* Answering
Knowledge* Extraction* Annotators
Knowledge* Validation* and*
Extrapolation
Language* Modeling
Latent* Semantic* Analysis
Learn* To*Rank
Linguistic* Analysis
Logical* Reasoning* Analysis
Logistical* Regression
Machine* Learning
Multi,Dimensional* Clustering
Multilingual* training
n,Gram* Analysis* (word*
combinations* and* distance)
Ontology* Analysis
Pareto* Analysis
Passage* Answering
PDF*Conversion
Phoneme* Aggregation
Question* Analysis
Question,answering* Reasoning*
Strategies
Recursive* Neural* Networks
Rules* Processing
Scalable* Search
Similarity*Analytics
Statistical* Language* Parsing
Support* Vector* Machines
Syllable* Analysis
Table* Answering
Visual* Analysis
Visual* Rendering
Voice* Synthesis
These*APIs*are*underpinned*by*
50#technologies:
2011
2015Source:*http://www.ibm.com/smarterplanet/us/en/ibmwatson/developercloud/services,catalog.html
Cognitive Services
Live Workshop
Chewing(
Gum(Wall(in(
California
Source:(http://en.geourdu.co/buzz/bizarre5shocking/chewing5gum5wall5in5california/
San(Luis(Obispo
Customer Analysis Healthcare
IBM Chef Watson.
Inspire your cooking decisions
Cognitive)
Cooking
187
Cognitive)Computing)approach)to)Computational)Creativity
Create&Food&new&
recipes&from&scratch
Modify&existing&rec
ipes&to&satisfy&your&
own&taste
Suggest&new&things&to&
prepare&&&cook
Pair(
ingredients(
and(flavors(for(
recipes(and(
dishes(
1876
Wisdom for All Nutrition
17. Fintech
Financial technology, also known as FinTech, is
an economic industry composed of companies
that use technology to make financial services
more efficient.
Financial technology companies are generally
startups founded with the purpose of disrupting
incumbent financial systems and
corporations that rely less on software.
Source: https://en.wikipedia.org/wiki/Financial_technology
17
19. « UNBUNDLING » is general phenomena that is impacting every sectors
or corporations
Source: https://www.cbinsights.com/blog/smart-home-market-map-company-list/
SMART HOMES
29. 29
What is the behind?
Digital Business + Digital Intelligence
30. 30
You shared
your position
with me and
can guess
your mobility
need.
I can take you
where you
need to be
Just enjoy
your new
experience
. Stay safe
as in your
home
I know
what is
needed for
you, even
before you
order it
Please, come
with me and
stay by me.
I know your
content I can
take care of all
your digital life
35. 78% of
Executives say
business will
manage
people along
side
machines
36.
37.
38.
39. 39
What is the behind?
Digital Business + Digital Intelligence
40. 40
You shared
your position
with me and
can guess
your mobility
need.
I can take you
where you
need to be
Just enjoy
your new
experience
. Stay safe
as in your
home
I know
what is
needed for
you, even
before you
order it
Please, come
with me and
stay by me.
I know your
content I can
take care of all
your digital life
43. 43
Image
source:
http://personalexcellence.co/blog/ideal-‐beauty/
City
Lifestyle
ZIPcode
Costal
vs
Inland Marital
status
Generation
Location
Family
Size
Gender
Income
Level
Competitors
Age
Loyalty
&
Card
Activity
Revenue
Size
Life
Stages
Eductation
Legal
status
Sector
Industry
Subscriptions
Date on Site
Wish List
Size of
Network
Check-ins
App usage duration
Number of Apps on Device
Deposits/Withdrawals
Device Usage
Purchase History
Following
Followers
Likes
Number of Hashtags used
History of Hashtags
Search Strings entered
Sequence of visits
Time/Day log in
Time spent on site
Time spent on page
Frequency of Search
Videos Viewed
Photos liked
44. 44
Image
source:
http://personalexcellence.co/blog/ideal-‐beauty/
City
Lifestyle
ZIPcode
Costal
vs
Inland Marital
status
Generation
Location
Family
Size
Gender
Income
Level
Competitors
Age
Loyalty
&
Card
Activity
Revenue
Size
Life
Stages
Eductation
Legal
status
Sector
Industry
Subscriptions
Date on Site
Wish List
Size of
Network
Check-ins
App usage duration
Number of Apps on Device
Deposits/Withdrawals
Device Usage
Purchase History
Following
Followers
Likes
Number of Hashtags used
History of Hashtags
Search Strings entered
Sequence of visits
Time/Day log in
Time spent on site
Time spent on page
Frequency of Search
Videos Viewed
Photos liked
Sentiment
Tone
Euphemisms
Hedonism
Extroversion
Face Recognition
Openess
Colloquialism
Reasoning Strategies
Language Modeling
Dialog
Intent
Latent Semantic Analysis
Phonemes
Ontology Analysis
Linguistics
Image Tags
Question Analysis
Self-transcendent
Affective Status
52. 52
BIG DATA
DATA
WISDOM
Knowledge
Information
Technology is no more
supporting every kind of
private and public
organizations, it is becoming
part of them.
Machine Intelligence
Is becoming the key
ingredient.
Analytics
Cloud Computing
Data Science
Mobile
Social
Digitalization
Technology
Business
Robotics
Artificial Intelligence
Business & Tech NexusThings
59. @pieroleo
59
Social
Data from and about People
Physical
Sensors & Streams
Terabytes to exabytes of
existing data
to process
Streaming data,
milliseconds to seconds to
respond
Structured, Semi-
structured Unstructured,
text & multimedia
Uncertainty from
inconsistency,
ambiguities, etc.
Volume
Velocity
Variety
Veracity
Data
Content
>80%
<20%
Traditional
Enterprise Data
Big data embodies new data characteristics created
by today’s digitized marketplace
Biological
DNA Sequencers
60. @pieroleo
60 60
Global Data Volume in Exabytes
Multiple sources: IDC,Cisco
100
90
80
70
60
50
40
30
20
10
Aggregate Uncertainty %
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
2005 2010 2015
By 2015, 80% of all available data will be uncertain: Veracity
Data quality solutions exist for
enterprise data like customer,
product, and address data, but
this is only a fraction of the
total enterprise data.
By 2015 the number of networked devices will
be double the entire global population. All
sensor data has uncertainty.
The total number of social media
accounts exceeds the entire global
population. This data is highly uncertain
in both its expression and content.
61. @pieroleo
Paradigm shifts enabled by big data and analytics
TRADITIONAL APPROACH
Analyze small subsets
of information
Analyzed
information
All
available
information
BIG DATA & ANALYTICS APPROACH
Analyze
all information
All
available
information
analyzed
Leverage more of the data being captured
Data leads the way— discover new emerging
properties
Reduce effort required to leverage data
Leverage data as it is captured
TRADITIONAL APPROACH
Carefully cleanse information
before any analysis
Small amount of
carefully organized
information
BIG DATA & ANALYTICS APPROACH
Analyze information as is,
cleanse as needed
Large
amount
of messy
information
Hypothesis Question
DataAnswer
TRADITIONAL APPROACH
Start with hypothesis and
test against selected data
BIG DATA & ANALYTICS APPROACH
Explore all data and
identify correlations
Data Exploration
CorrelationInsight
Repository InsightAnalysisData
TRADITIONAL APPROACH
Analyze data after it’s been processed
and landed in a warehouse or mart
Data
Insight
Analysis
BIG DATA & ANALYTICS APPROACH
Analyze data in motion as it’s
generated, in real-time
63. @pieroleo
Just ONE Transaction
path goes to the end in
thousands and to
complete that path tens
of decision points were
considered. Right now
we store and analyze in
our transactional
systems just the
transaction end points.
Buyer ….Win!!!
Buying Decision Labyrinth
Yes!
Big
Data
is
the
answer
and
the
need
of
the
new
emerging
sub-‐
transactional
era
64. @pieroleo
It's an invitation-only loan product offered exclusively to Amazon Sellers. The Amazon
loans offers very competitive from 6 to 14% interest rates and no pre-payment penalty.
The power of a sub-transactional knowledge
Source: http://uk.businessinsider.com/r-exclusive-amazon-to-offer-loans-to-sellers-in-china-7-other-countries-
2015-6?r=US&IR=T
US, Japan from 2012 and from 2015 - Canada, China, France, Germany, India, Italy, Spain and the United Kingdom
66. @pieroleo
Source: Cornell University - Maize kernal infected with Aspergillus flavus, which produced
aflatoxin.http://www.plantpath.cornell.edu/labs/milgroom/Research_aflatoxin.html And http://www.special-clean.com/special-
clean/en/mold/mold-lexicon-1.php
For science, Big Data is the microscope of the 21st century
67. @pieroleo
Source: A statue representing Janus Bifrons in the Vatican Museums
Big Data as a new Business Concept and as a new
Technology Concept
68. @pieroleo
68
Big Data as a new business concept:
New values and opportunities for a number of stakeholders
Chief Marketing Officer
how to improve customer focus?...could predict the right offer
for the right customer at the right time and improve customer
value and intimacy or prevent churn?
Chief Product Designer
...how we can innovste? … could
we improve our product
channels/design offering??
Chief Finance
Officer
...could streamline
compliance and
understand risk
exposure across
businesses and
regions?
Chief Risk Officer
...uses anti fraud predictive analytics to detect and
prevent rapid fire anomalous transactions or wire
transfers identified as high probability of fraud?
Chief Executive Officer
...could make better business decisions
using accurate data across all
company/system dimensions and
across time horizons: past, present and
future?
Chief Information Officer
...could analyze oceans of machine generated logs to
predict which components or equipment in the
datacenter are likely to fail and thereby avert a disruption
during critical quarter end? How we can support Zero
high risks or manage crisis?
Big
Data
69. @pieroleo
We need to combine internal and external data, utilized and under-utilized data,
structured and unstructured data... and cross-link organization knowledge & data
silos
CRM
• emails
• claims
• call center scripts
• Chats with customers
• …
Transactional Info.:
• Transactions
• Orders
• consultancies
• …
Legal Info:
• Contracts
• Complaints
• Reports
• Legal Actions
• Fraud Data
• …
Knowledge Management
•Manuals, wikis, couses
•Projects Data
•Market Analysis
•RSS Business Feeds
•Data feed: Bloomberg reuters
• …
IT Systems
System Logs
Application logs: web, vending machines,
mobile
Video
Sensor Networks, RFID
• …
Social Media:
• Global Social Networks: tweeter,
facebook, etc.
• Small communities: blogs, muros
corporativos,
• Internal Social Networks
(employees)
• News
• … Big
Data
Big Data as a new technology concept
70. @pieroleo
“Big Data is the set
of technical
capabilities,
management
processes and
skills for converting
vast, fast, and varied
data into Right Data
to produce useful
knowledge”
Source:
Definition discussed during the work of the
Word Summit on Big Data and Organization
Design Paris – 2013 and Adapted from:
Beacon Report – Big Data Big Brains – 2013
In summary, what is Big Data?
71. @pieroleo
New Organization Design: What is New and Different?
A lot more data and different
kinds of data.
Historically most data was structured data – rows and
columns
Today it is unstructured data like aerial photos, audio
from call centers, video from surveillance cameras, e-
mails, texts, diagrams.
A shift in focus from data
stocks to data flows.
Historical information was stored in data warehouses
and analyzed by data mining.
Streaming data arrives in real time allowing us to
influence events as they happen. We can prevent some
bad events from ever happeningat all.
Shift in the power structure of the
company. Many companies have analog
establishments. We need to shift power to
those who can draw valuable insights from
data and analytics and implement them.
Shift from periodic to real time or
continuous decision making. We need an
increase in the clock speed of every process
in the company.
There is a potential for “Big Data” to
become a fundamental center for the
company. Is it a new dimension of
structure?
Organization Design IssuesTechnology Issues
Source: Jay R. Galbraith
75. Toward a Precise Decision Making to reduce the wasteful
spend as well as the risk in every industry
New Information Technology challenge is now about the
possibility to expand our WISDOM options
Watson
76. 2011
2015
2016 - AlphaGO=4 Lee Se-Dol=1
1997 - IBM=2.5 Kasparov=2.5
1997
AlphaGO uses self-trained net to evaluate
positions and moves on 30M historical
games
DeepBlue uses a hard-coded objective function
written by a human coupled with High
Performance Computing
2016
10
10170
1040
Applying or having wisdom in real world is
not only an AI game
COMPUTING & MATH WISDOM
IBM Watson – Jeopardy!
SEMANTICS
77. The Jeopardy! Challenge: 5 Key Dimensions to drive
Question Answering
Broad/Open
Domain
Complex
Language
High
Precision
Accurate
Confidence
High Speed
$600
In cell division, mitosis
splits the nucleus &
cytokinesis splits this
liquid cushioning the
nucleus
$200
If you're standing, it's the
direction you should look
to check out the
wainscoting.
$2000
Of the 4 countries in the
world that the U.S. does
not have diplomatic
relations with, the one
that’s farthest north
$1000
The first person
mentioned by name in
‘The Man in the Iron Mask’
is this hero of a previous
book by the same author.
What is down?
Who is
D’Artagnan?
What is
cytoplasm?
What is North
Korea?
81. 81
Analytic
Systems use
statistical
techniques for
detecting patterns
or detect trends
within data, yield
an understanding
of historical or
current state from
which to draw
conclusions
Text Mining is a class of functions for
parsing and identifying significant words
in language (NLP) as well as
understand the semantic of a textual
content
Cognitive Systems
leverage machine
learning to predict
meaning in features of
human language
(spoken, written, visual)
and related forms of
human reasoning
Multi-Media Mining is a a class of
function for analyzing visual content
such as images or videos
Speech Mining is a class of
functions for analyzing audio
signals including speech to such
as ability Cognitive Solutions
leverage a combination of
cognitive system reasoning
strategies and other analytic and
classical computing techniques to
solve for a complex problem ->
Amplify Human WISDOM in a
specific domain
XXX Mining is class of large
specialized functions for analyzing
“digital representation” in a
specific domain à e.g.,
Bioinformatics, Financial Analytics,
etc.
Machine Learning
is a class of
statistical techniques
that use training
data to recognize
the correlation
between a set of
feature patterns and
outcomes.
It includes also Deep
Learning that is a
rapidly maturing space,
based on neural
network techniques,
that are taught to find
their own features
Emerging Patterns for Artificial Intelligence adoption in Business World
WISDOMBIG DATA ANALYTICS
82. @pieroleo
82
• Cognitive systems are able to learn their behavior through
education;;
• That support forms of expression that are more natural for
human interaction;;
• Whose primary value is their expertise;; and
• That continue to evolve their reasoning approach as they
experience new information, new scenarios, and new
responses
1.education 2.expression 3.expertise 4.evolve
Which are cognitive systems main attributes?
86. Ecosystem
and Partners
Industry
Solutions
Client
Solutions
& products
IBM
Provided
Data Publically
Sourced
Data
Partner
Provided
Data
Private
Client
Data
IBM Watson Innovation platform for Cognitive Business
Watson Health
Watson Financial Service
Watson Internet of Things
Hybrid Watson
Frameworks
Watson
Services - API
Data
Knowledge
Wisdom
88. Ecosystem
and Partners
Industry
Solutions
Client
Solutions
IBM
Provided
Data Publically
Sourced
Data
Partner
Provided
Data
Private
Client
Data
IBM Watson Innovation platform for Cognitive Business
HealthFinancial
Cross
Public Filings
Patents
Medical Journals
U.S. Geological Survey
…
Apple
Twitter
Quest Diagnostics
…
Medtronic
Under Armour
Johnson & Johnson
Thomson Reuters
…
Watson Health
Watson Financial Service
Watson Internet of Things
Hybrid Watson
Frameworks
Watson
Services
Comms Industrial Distribution Financial Public ServicesHealth
Fraud Analysis
Corp Intelligence
Claims Processing
Digital Agent
Call Center Advisor
Public Safety
National Security
Shopping Advisor
Sales Automation
Supply & Logistics
Omni-Channel Ops
Product Safety
Field Service Mgt
Geology Advisor
Digital Agent
Theme Park Exp
Call Center Ops
CIO Dashboard
Corp Intelligence
M&A Advisor
Cyber Security
Life
Sciences
Oncology
Clinical Trial
Matching
89.
90.
91.
92.
93. 1-800 Flowers
Live at: https://www.1800flowers.com/gwyn-1800flowers?flws_rd=1 Live at: https://www.thenorthface.com/xps
GWYN (Gifts When You Need), a Watson-powered personal
concierge designed to help customers find the perfect gift
The North Face
A personal Shop Assistant that can drive you to select the
most appropriate Jacket
Virtual Agents: Sales Assistants
94. • Will deliver personalized content
through the dashboard and other
digital channels supported by the
OnStar Go ecosystem to make the
most of time spent in the car.
• iHeartRadio will use Watson
Personality Insights to curate
personalized experiences that
leverage on-air personalities and
local content from radio stations
across the U.S.
• The platform employs Watson
Tradeoff Analytics to give a traveling
foodie dining recommendations
from celebrity chefs when driving in
a new city.
Cognitive Automation
98. 98
8,361Teams joined to propose and generate ideas
And over 2.700 passed feasibility reviews
275,000 IBMers all around the world who engaged in the Cognitive Build.
• Imagine a digital cognitive system to help you do something important in your
personal or professional lives
• Team to design it and advocate for it, and then everyone votes
• Winners: reduce waste and human suffering, screen for health issues and safety
threats, learn life skills and make better choices, find what you are looking for,
move around more effectively, provide emotional support, provide IT support,
learn about important public policy goals and make better choices
99. Types of Cognitive Systems
99
Tool AssistantTools Collaborator
Coach Mediator
Source: Analysis of top 400 ieas by J. Spoorer, Don Norman and Paul Maglio
101. Ecosystem
and Partners
Industry
Solutions
Client
Solutions
IBM
Provided
Data Publically
Sourced
Data
Partner
Provided
Data
Private
Client
Data
IBM Watson Innovation platform for Cognitive Business
Watson Health
Watson Financial Service
Watson Internet of Things
Hybrid Watson
Frameworks
Watson
Services
Data
Knowledge
Wisdom
103. Anaphoric Co-referencing
Colloquialism Processing
Content Management -- Versioning
Convolutional Neural Networks
Curation
Deep Learning
Dialog Framing
Ellipses
Embedded Table Processing
Ensembles and Fusion
Entity Resolution
Factoid Answering
Feature Engineering
Feature Normalization
Focus and Spurious Phrase
Resolution
HTML Page Analysis
Image Management
Information Retrieval
Knowledge (Property) Graphs
Knowledge Answering
Knowledge Extraction Annotators
Knowledge Validation and
Extrapolation
Language Modeling
Latent Semantic Analysis
Learn To Rank
Linguistic Analysis
Logical Reasoning Analysis
Logistical Regression
Machine Learning
Multi-Dimensional Clustering
Multilingual training
n-Gram Analysis (word
combinations and distance)
Ontology Analysis
Pareto Analysis
Passage Answering
PDF Conversion
Phoneme Aggregation
Question Analysis
Question-answering Reasoning
Strategies
Recursive Neural Networks
Rules Processing
Scalable Search
Similarity Analytics
Statistical Language Parsing
Support Vector Machines
Syllable Analysis
Table Answering
Visual Analysis
Visual Rendering
Voice Synthesis
These APIs are underpinned by
50 technologies:
2011
2015Source: http://www.ibm.com/smarterplanet/us/en/ibmwatson/developercloud/services-catalog.html
104. IBM Cognitive Services
1. Watson APIs are
continuously.
2. They are
complemented with
tens of other APIs in
other domains, all
running on ONE
platform.
3. They can mashed
up to build an
infinite number of
cognitive assistants.
2011
2016
Pipeline
105. Gain insight into how
and why people think,
act, and feel the way
they do. This service
applies linguistic
analytics and
personality theory to
infer attributes from a
person's unstructured
text
Personality
Insights
108. Source: https://ibmtjbot.github.io/
I'm an open source project
designed to help you access
Watson Services in a fun way.
You can 3D print me or laser
cut me, then use one of my
recipes to bring me to life!
https://www.ibm.com/watson/developercloud/project-intu.html
112. @pieroleo
Understands the language of business
Visual, simple and
intuitive
Simply type in a
question and get
meaningful
insights
immediately
Visual, simple and
intuitive
Automatically
suggests graphs and
visuals to
communicate
findings
INSIGHTContext
Automatically
presents related facts
and insights to guide
discovery
insight
insight
insight
insight
insight
insight
insight
You and your business data
https://www.analyticszone.com/homepage/web/displayNeoPage.action
113. Even a simple analytics project has multiple
steps and people
Data
Access
Data
Preparation
Analysis
Validation
Collaboration
Reporting
Data Scientists
and Statisticians
Business
Users
IT
Business
Analysts
114. And it’s rarely a straightforward process
Data
Access
Data
Preparation
Analysis
Validation
Collaboration
Reporting
Data Scientists
and StatisticiansBusiness Users
IT
Business
Analysts
118. Single Interface … Explore > Predict > Assemble
Quick start
intuitive
interface
Key business
driver insights
Dashboard
and
storytelling
authoring
Natural
language
dialogue
Easy data
upload and
Refinement
capabilities
119.
120.
121. @pieroleo
IBM Watson Analytics
Watson Analytics
Communication & Collaboration
Visualization & Storytelling
Analytics
Descriptive, Diagnostic, Predictive, Prescriptive, Cognitive
Data Access & Refinement
Cloud
Operations H
R
ITFinanceSalesMarketing
Mobile Ready Secure
Value:
•Put analytics in the hands of
everyone
•Make access to data easy for
refinement and use
•Deliver through the cloud for agility
and speed
Prioritizing
Accounts
Receivable
Identifying
and
Retaining
Key
Employees
Helpdesk
Case
Analysis
Campaign
Planning
and
ROI
Warranty
Analysis
Customer
Retention
Finance HRITMarketing OperationsSales
Examles
123. 123
Analytic
Systems use
statistical
techniques for
detecting patterns
or detect trends
within data, yield
an understanding
of historical or
current state from
which to draw
conclusions
Text Mining is a class of functions for
parsing and identifying significant words
in language (NLP) as well as
understand the semantic of a textual
content
Cognitive Systems
leverage machine
learning to predict
meaning in features of
human language
(spoken, written, visual)
and related forms of
human reasoning
Multi-Media Mining is a a class of
function for analyzing visual content
such as images or videos
Speech Mining is a class of
functions for analyzing audio
signals including speech to such
as ability Cognitive Solutions
leverage a combination of
cognitive system reasoning
strategies and other analytic and
classical computing techniques to
solve for a complex problem ->
Amplify Human WISDOM in a
specific domain
XXX Mining is class of large
specialized functions for analyzing
“digital representation” in a
specific domain à e.g.,
Bioinformatics, Financial Analytics,
etc.
Machine Learning
is a class of
statistical techniques
that use training
data to recognize
the correlation
between a set of
feature patterns and
outcomes.
It includes also Deep
Learning that is a
rapidly maturing space,
based on neural
network techniques,
that are taught to find
their own features
Emerging Patterns for Artificial Intelligence adoption in Business World
WISDOMBIG DATA ANALYTICS
124. Massive Unstructured is the biggest data wave of all
1990’s 2020’s
Video
Text
Exa
Peta
Tera
Giga
Data Volume
2000’s 2010’s
Structured data
Audio
Image
Med
High
Low
Computational Needs
Sophistication of Analysis
Expressiveness
Digital Marketing
10+% of video views
Wide Area Imagery
100’s TB per day72 video hrs/minute
Media
Source: IBM Market
Insights based on
composite sources
Safety / Security
Healthcare
Customer
1B camera
phones
1B medical images/yr
10s millions cameras
Enterprise Video
Used by 1/3 of
enterprises
125. Structured versus Unstructured Information: What
does it mean?
Know this is the last name and this is their age
The information is unambiguous
The context of the information is known
Pre-defined and
machine-
readable
126. Structured versus
Unstructured Information: What does it
mean?
Office Location is unstructured
Address
City
Zip code
….
127.
128. The Enquire reported that the attractive, Ms Brown,
CEO of Textract Corp, had been recently spotted drunk at
Summit meeting in Zurich,…………At 42, Ms. Brown, is
the youngest CEO at the Summit,…
<Organization>
<Name>
<Title>
<Proper Name> <Occupation>
Example of Annotation of a Text – “construct meaning from
free form text, include identification and labeling the text
with specific meanings”
<Positive ><Negative >
Unstructured Information:
The context of the information is not known and is interpreted by the
computer using mathematical techniques
129. Text Mining: transforms
UnStructured Information into Structured data
Before After
Concept/entity extraction
Relationship extraction
Sentiment Analysis
Linguistic Analysis
Categorization
Clustering,
Text Analytics
Tasks
Document
Summarization
….
130. Automotive Quality Insight
• Analyzing: Tech notes, call logs, online
media
• For: Warranty Analysis, Quality Assurance
• Benefits: Reduce warranty costs, improve
customer satisfaction, marketing
campaigns
Crime Analytics
•Analyzing: Case files, police records, 911 calls…
•For: Rapid crime solving & crime trend analysis
•Benefits: Safer communities & optimized force
deployment
Healthcare Analytics
• Analyzing: E-Medical records, hospital
reports
• For: Clinical analysis;; treatment protocol
optimization
• Benefits: Better management of chronic
diseases;; optimized drug formularies;;
improved patient outcomes
Insurance Fraud
•Analyzing: Insurance claims
•For: Detecting Fraudulent activity &
patterns
•Benefits: Reduced losses, faster
detection, more efficient claims processes
Customer Care
• Analyzing: Call center logs, emails, online
media
• For: Buyer Behavior, Churn prediction
• Benefits: Improve Customer satisfaction
and retention, marketing campaigns, find
new revenue opportunities, recostruct life
stages and life events
Social Media for Marketing
• Analyzing: Call center notes, multiple
content repositories
• For: churn prediction, product/brand
quality
• Benefits: Improve consumer satisfaction,
marketing campaigns, find new revenue
opportunities or product/brand quality
issues
A first set of examples
leveraging Text Mining / Analytics
132. Multimedia Mining flow: Feature extraction, modeling, and
application of semantics and context are required to deliver
insights
Labeled DataUnlabeled Data
K-means Bayes NetClustering
Markov
Model
Decision
Tree
Modeling
Color
Spectrum
Edges
Camera
Motion
Feature Extraction
Ensemble
Classifiers
Texture
Active
Learning
Deep
Belief Nets
Vehicle tracking Activity classificationSafe zone monitoring
Locations Activitie
s
Scenes
Safety/Security
Behaviors
Objects
PeopleEvents
Tracks
Moving
Objects
Actions
Neural
Net
classification
scoringSemantics
Multimedia
AdaBoost
Blobs
Background
Segmentation
Zero-crossings
Support
Vector Machine
Gaussian
Mixture Model
Hidden
Markov
Model
Frequencies
133. Video-based Appraisal:
§ Goal: improve home, automobile,
or marine insurance process using
supporting multimedia data
§ Use video by insurance policy
holder to document insured items
§ Automatically turns the video into
the basis for appraisals and claims
Insurance
Public Safety and Security:
§ Goal: ensure safety and security
in transit system
§ Detect suspicious activities, safety
concerns, and crowd conditions
using camera-based analytics
§ Support real-time alerting and forensic
search over video data
Transportation
In Store Video Analytics:
§ Goal: use existing store cameras
to tell who is entering the store and
demographics
§ Bring video to aisles to tell how long
people look at products and ads, what
they picked up, whether they placed in cart
§ Extend campaign management and customer
analytics solutions with in-store analytics
Retail
Consumer Goods
Identify Logo Exposure:
§ Goal: automatically annotate
videos with logo version and
calculate exposure time
§ Identify multiple logo appearances
in the same frames
§ Identify distorted logos on clothing
and promotional items
Many enterprises are investigating next
generation multimedia analytics-based solutions
135. Chewing
Gum Wall in
California
Source: http://en.geourdu.co/buzz/bizarre-shocking/chewing-gum-wall-in-california/
San Luis Obispo
136. Portraits from New York
Stranger
Visions
In Stranger Visions artist Heather Dewey-Hagborg creates portrait sculptures from analyses of DNA
material collected in public places.
Source: http://deweyhagborg.com/strangervisions/
137. Customer Analytics: Adding Value at Every Point of Interaction
and leveraging customer Digital Footprints
Systems of Record Systems of
Engagement
Customer
Analytics
Big Data Analytics
138. 138
All perspectives
Past (historical, aggregated)
Present (real-time, scenarios)
Future (predictive,
prescriptive)
At the point
of impact
All decisions
Major and minor;;
Strategic and tactical;;
Routine and exceptions;;
Manual and automated
All information
Transaction/POS data
Social data
Click streams
Surveys
Enterprise content
External data (competitive,
environmental, etc.)
All people
All departments
Front line, back office
Executives, managers
Employees
Suppliers, customers and
consumers
Partners Customer
Analytics
Challenge: Consider all data points
139. What
are
people
saying?
How
do
people
feel
about
my
brand?
Who
is
this
individual
like?
Who
does
she
influence/follow?
What
are
her
preferences?
What
words/offers
will
engage
her?
Customer Analytics
Practical CHALLENGES
141. SINGLE
VIEW
Business
Data,
Social
Data,
Interactive
data
360°Integrated
Customer View
Marketing
Cust. Care
Sales
Risk, Fraud
Customer Analytics challenge:
build a 360°Integrated Customer View
… and more
142. SINGLE
VIEW
Business
Data,
Social
Data,
Interactive
data
360°Integrated
Customer View
Marketing
Cust. Care
Sales
Risk, Fraud
How?Why?
Who? What?
Customer Analytics challenge:
build a 360°Integrated Customer View
… and more
143. Social Data is not a SINGLE and omogeneos source: it is a complex aggregate of content that
we can leverage in dependance of well defined Business Use Cases.
General Rule for Social Data
144. Examples of Social Media Outlets
§ More than 1 billion unique users visit Youtube each
month watching over 6 billion hours of video
§ More than 388 million people view more than 12.7
billion blog pages each month
§ There are 500 million tweets daily – that’s 5,700 per
second
§ 50% of Facebook users check it daily – there are
more than 1 billion users world wide
1
145. Monitoring
and
Reporting
Analytics
of
Aggregates
Analytics
of
Individuals
&
specific
groups
Listening
Engagement
Demographics
Publishing
Measurement Net
Promoter
Network
Topology
Sentiment
Analysis
Brand
Analysis
Identity
AnalysisPredictive
Analysis
SNA Pattern
Detection
Intrinsic
Preferences
Social
GenomeMicro-‐Segmentation
Next
Best
OfferMessaging/campaigns
Face
Recognition
Visual
Recognition
Age
Detection
Image
Tagging
Gender
Recognition
Identity
Recognition
What
are
people
saying?
How
do
people
feel
about
my
brand?
Who is this individual like?
Who does she influence/follow?
What are her preferences?
What words/offers will engage her?
Techniques
Cognos - Big Insights – SMA - SPSS –
Watson Explorer – Adv. Analytics & Cognitive Services
From CHALLENGES to Techniques
And Capabilities
148. CustomerAnalytics &
TRUST
“Trust men and they will be true
to you;; treat them greatly and
they will show themselves
great.”
Ralph Waldo
Emerson
149. Consumers are open to share their personal information,
with the exception of financial data, when there is
perceived benefit
Consumer Maintains Control of Data
What is your willingness to provide information in exchange
for something relevant to you (non-monetary)?
Source:
IBV
Retail
2012
Winning
Over
the
Empowered
Consumer
Study
n=
28527
(global)
P04:
What
is
your
willingness
to
provide
information
for
each
of
the
following
items
if
[pipe
primary
retailer]
provided
something
relevant
to
you
in
exchange?
25% 27%
41% 41% 44% 46%
63%
30% 30%
28% 29% 28% 28%
21%
45% 43%
33% 30% 28% 26%
15%
0%
20%
40%
60%
80%
100%
Media Usage
(e.g. Media
channels)
Demographic
(e.g. age,
ethnicity)
Identification
(name,
address)
Lifestyle (# of
cars, home
ownership)
Location
Based
Medical Financial
Completely Disagree Neutral Completely willing
159. @pieroleo
170,000 personal weather
stations worldwide
2.2 B locations forecasted every
15 minutes.
15 B Weather averages 15B
forecast queries daily.
20 terabytes, every day.
Bring Advanced Weather Insights to Business
Source: https://www.wunderground.com/
165. Ecosystem
and Partners
Industry
Solutions
Client
Solutions
IBM
Provided
Data Publically
Sourced
Data
Partner
Provided
Data
Private
Client
Data
IBM Watson Innovation platform for Cognitive Business
Watson Health
Watson Financial Service
Watson Internet of Things
Hybrid Watson
Frameworks
Watson
Services - API
Data
Knowledge
Wisdom
166. Leveraging the Explosion of Data in Medicine – An Impossible Task Without
Analytics and New advanced Artificial Intelligence Computing Models
1000
Facts
per
Decision
10
100
1990 2000 2010 2020
Human
Cognitive
Capacity
Electronic
Health
Records
(Clinical
Data)
Internet
of
Things
(Exogenous
Data)
The
Human
Genome
(Genomic
Data)
Capturing
the
Value
of
Data:
Big
Changes
Ahead
Medical error—the third leading cause of
death in the US
Source:
BMJ 2016;; 353 doi:
http://dx.doi.org/10.1136/bmj.i2139 (Published 03 May
2016) Cite this as: BMJ 2016;;353:i2139
167. Ecosystem
and Partners
Industry
Solutions
Client
Solutions
IBM
Provided
Data Publically
Sourced
Data
Partner
Provided
Data
Private
Client
Data
An example of industrial-oriented platform: Watson Health
Watson Health
Watson Financial Service
Watson Internet of Things
Data
Knowledge
Wisdom
Public Filings
Patents
Medical Journals
Apple
Twitter
Quest Diagnostics
Medtronic
Under Armour
Johnson & Johnson
TEVA
168. 168
Watson Health is bringing unique insights to the marketplace to help reduce
costs, improve outcomes and help increase value.
Data
Standards
based,
extremely
scalable,
open
repository
of
data
on
all
dimensions
of
healthcare
and
research
Insights
as
a
service
Knowledge and actionable
information through
advanced analytics and
cognitive capabilities
Solutions
IBM and an ecosystem
of partners help improve
the overall experience
and increase the quality
of outcomes
Watson Health
Data – Insights – Solutions
170. Watson Health’s aim is to create an open industry platform utilizing key
capabilities and partnerships to help improve Healthcare
Watson
Cloud
PARTNERSHIPS
172. Watson for Genomics
Business Challenge:
• As the cost of Next Generation Sequencing decreases, there will be an increase in tumor
genome sequencing resulting in massive quantities of genetic data to analyze
• Currently, it takes an average of 4-6 weeks to analyze and interpret genetic data manually
• Complexity of matching genetic mutations of individual’s tumor with molecular targeted
therapies using multiple data sources
Watson Solution:
• Empowers Physicians to Make the Most of Genomic Data and Assisting Them to Provide
Comprehensive and Up-to-date Cancer Patient-Care
1. Leverages whole genome, whole exome, or large panels variant sequences from patient tumor biopsies
2. Identifies gene level variants using several industry standard databases, as well as relevant literature
3. Provides actionable list of gene variants and the therapies that target them, either directly or indirectly
Use Cases:
• Assist Molecular Pathologists in reviewing the 100s to 1000s of gene level variants, and
associating each with the likelihood its driving cancer developing in that individual patient
• Once the driver alterations have been approved by the pathologist, WGA assists the Medical
Oncologist with recommending an approved, investigational, or off-labeled targeted therapy
172
173. Watson Genomics from Quest Diagnostics®
Watson Genomics from Quest Diagnostics is a solution
that can help patients along their cancer journey.
1. Quest Diagnostics sequences and analyzes a tumor’s
genomic makeup to find specific mutations
2. Watson then compares those mutations against
relevant medical literature, clinical studies,
pharmacopeia and carefully annotated rules created
by leading oncologists.
3. A Quest Diagnostics pathologist will review and
validate the results and prepare a report to send back
to the patient’s treating physician
http://www.ibm.com/watson/health/oncology/genomics/
179. 179
The
Medical
Sieve §Build a fast anomaly detection
engine
– Quickly filters irrelevant images
– Highlights disease-depicting regions
– Flags coincidental diagnosis
§ Intended as a radiology assistant
– Clinicians still do the diagnosis
– Machine reduces workload
– Machine performs triage/decision
support
Given history of the patient and images of
a study
Is there an anomalous image here?
If so, where is the anomaly ?
Describe the anomaly
The
Medical
Sieve
182. @pieroleo
182
Pathway
Genomics
OME
App
– Powered
by
Watson
Merging
cognitive
computing
and
deep
learning
with
precision
medicine
and
genetics
How it works
Pathway Genomics mails
the user a saliva DNA
collection kit
Pathway will work with clinicians
and scientists to conduct the
Pathway Fit test. It specifically
looks at 75 genes that focus on
phenotypes like diet, exercise,
lipids, and sugar metabolism
Watson cognitive computing
technology, intelligent machine
learning, and a corpus of
health and wellness
information
With Watson APIs, the
Pathway app leverages
Watson’s natural language
processing technology and
content in the form of health
and wellness information
Highly personalized insights to empower
people to change unhealthy behaviors,
allowing them to live healthier lives, e.g.
genetically optimal diet plans or
restaurant and menu recommendations
Early
Alpha
Version
Users
unique
genetic
traits Health Habits
Data
from
wearable
health
monitors
Apple HealthKit Electronic
health
records Insurance
informationGPS Data
Incorporated Data: Pathway’s “FIT” Test
Additional
datasets
Other User Data Watson corpus of health and wellness information
Data Sources
186. Food
Security
Cooking
Health
Wellbeing
Nutrition & Technology
AI & Machine
Learning
Digital Data
Cloud
Analytics
Agroindustry
Internet of
Things
Genomics
Metabolomics
Food
Distribution &
Preparation
There is a nexus of forces, from different angles, that combine
Nutrition & Technology
Creativity
Computing
An opportunity to support decisions of professionals and consumers
with data is emerging
Mobile
Social
3
188. IBM Chef Watson.
Inspire your cooking decisions
Cognitive
Cooking
188
Cognitive Computing approach to Computational Creativity
Create
Food
new
recipes
from
scratch
Modify
existing
rec
ipes
to
satisfy
your
own
taste
Suggest
new
things
to
prepare
&
cook
Pair
ingredients
and flavors for
recipes and
dishes
1886
193. https://twist.ibmchefwatson.com/
Tell
Watson
how
you
are
feeling
and
how
to
start
to
drink
Tweak
your
flavors
based
on
Wa
tson’s
analysis
and
suggestions
Bring
the
flavors
to
life
with
your
bart
ender,
snap
a
photo
and
share!
198. Weather is
the secret to understanding
how consumers feel… and cook
A brand able to gain a
spot in the daily
routinesand rituals of
consumerscreates a not
only a relationbut a
deep intimacy with
them
198
202. @pieroleo
Scientific
Method
Visualization
Domain
Expertise
TOM
Hacker
Mindset
MathData Engineering
Advanced
Computing
StatisticsData Scientist
A Data Scientist
§ Explores and examines data from
multiple disparate sources
§ Sifts through all incoming data with
the goal of discovering a previously
hidden insight
§ Has strong business acumen,
coupled with the ability to
communicate findings to both
business and IT leaders in a way that
can influence how an organization
approaches a business challenge
§ Represents an evolution from the
business or data analyst role
§ Has a solid foundation typically in
computer science and applications,
modeling, statistics, analytics and
math.
The role of a Data Scientist
206. @pieroleo
Sheryl Sandberg, COO, apologised for 'poor
communication' of the study
Said Facebook never meant to upset users with the
secret research
Was part of a study to see if people's moods are
affected by content
Information Commissioner now investigating whether
or not the site breached data regulations
Facebook has apologised to its
users after a secret psychological
experiment has sparked outrage in
the online community
Facebook admitted it
had manipulated the
news feeds of nearly
700,000 users
without their
knowledge as part of
a psychology
experiment.
Source: http://www.forbes.com/sites/kashmirhill/2014/07/02/sheryl-
sandberg-apologizes-for-facebook-emotion-manipulation-study-kind-of/
With Big Data #TRUST (plus #Security
plus #Privacy) matter
207. @pieroleo
“…Unfortunately, the conversations didn't stay
playful for long. Pretty soon after Tay launched,
people starting tweeting the bot with all sorts of
misogynistic, racist, and Donald Trumpist
remarks. And Tay — being essentially a robot
parrot with an internet connection — started
repeating these sentiments back to users,
proving correct that old programming adage:
flaming garbage pile in, flaming garbage pile
….“out.