I delivered this presentation at University at Chieti-Pescara in Abruzzo (Italy) in September 2015, introducing IBM Academy of Technology and talking about Cognitiva Computing and Analytics with IBM Watson and IBM IT Operations Analytics Log Analysis (ITOA). The video in Italian is available on YouTube, please contact me if you are interested. Thanks to Amanda Tenedini for the help with Social Media and to Piero Leo for the help with IBM Watson.
10. We maintain continuity in 2014 by carrying over 7 themes which will be “explanded” from
last year and introducing three new themes: Cognitive Business, New SW Models and IT
Infrastructure of Tomorrow
2013 2014
Analytics
Cloud Computing
Mobile Computing
User Experience
Integration by Design
Artificial Intelligence
Internet of Things
Social Media
Intelligent Automation
Digital Transformation
Wearable Computing
Information
Mngt &
Storage
Energy &
Green Comp.
Business
Applications
Platforms
Systems of
Engagement
Security and Privacy
Cognitive Business
New SW Models
IT Infrastructure of Tomorrow
2015 Top 10
Themes2012
Academy of Technology Top 10 Themes for 2015
• Look at #IBMAoT Twitter hashtag
14. Data growth and gravity distorts and impacts
every component of IT – and business
Has data a gravity?
15. Data at the edge is changing how we look at data
and how computing will evolve
90%
By 2017
Of data created over
the last 10 years was
never captured or
analyzed
The collective
computing and storage
capacity of
smartphones will
surpass all worldwide
servers
60%
2X
Of valuable sensory
data loses value in
milliseconds
Rate of data creation
compared to the
expansion of bandwidth
over the past decade
16. Source: "The Relative Contribution of Multiple Determinants to Health Outcomes", Lauren McGover et
al., Health Affairs, 33, no.2 (2014)
60%
30%
10%
Clinical data
Genomics data
Exogenous data
1,100 Terabytes
6 TB
0.4 TB
Per lifetime
Influence on outcomes
Example: Rapid Growth of exogenus data is transforming healthcare
17. Clinical data
Genomics data
Exogenous data
Glucose Monitoring
Calorie Intake
Stress Levels
Physical Activity
Other vital signs
Social Interaction
Affinity (retail)
Sleep Pattern
Employer (HR)
Disease management and
treatment
Behavioral change and
prevention
60%
Influence
on
outcomes
Leveraging Exogenous Data for Chronic Care (Type 2 Diabetes;
Primary & Secondary Prevention)
18. How to manage the new complexity?
Cognitive Computing is the answer
19. TIC
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 as they experience new information, new scenarios, and new
responses
Learn &
Improve
Speed &
Scale
Interact in a
natural way
Assist &
Augment
Cognitive computing principles
23. Passage What are the requirements for car insurance in Kentucky?
FAQs (aka TOA) What are your visiting hours?
Directed Can you walk me through installing a new app on my phone?
Actions I want to submit a claim!
Factoid Who is known as the father of modern psychology?
Definition What is an algebric structure?
Knowledge What is the capital of Texas?
Yes/No Do I need a visa to enter Brazil?
Query What is the highest priced phone offered by Samsung?
Calculated What is ½ Teaspoon of Tarragon in Grams?
Visual Who is this [person I have showing a picture of]?
State What is my current data usage?
Prediction What level of customer churn will we see next quarter?
Questions & Answers challenges as cognitive information
25. Develop hardware and
software systems with the
intent of reach and
superate human
intelligence
Develop systems able to
make us much more
productive!
Strong approach
Information-based Intelligence
Where Cognitive computing come from?
Artificial intelligence
26. First used approach, based on logical and mathematical programming,
thinking technologies, learning, planning and so on...
Many research efforts have been done until years ‘60 and ‘70 in several
worls Universities, thinking that the problem was the speed of computers
This made IT companies think that,when computers would have reached an
extremely high calculation and decision speed, then the result would have
arrived
This partially wrong believe was the cause of failures in the previous
millennium, although efforts of Japanish companies
Something more than speed was needed
Strong approach for Artificial Intelligence
27. Information Based Intelligence
Statistic methodologies, based to computational mathematical
calculation with very strong algoritms and high performances
Scalability, that means more information are available, more
powerful are computers, much more sophisticated can be
mathematical algoritms taking to better results
Knowledge, based on the huge quantity of information the
computer is able to ingest in order to produce valid
interpretation of data to pass for example to physical
experiments
• Evolution in the years
- Data mining around ’90
- IBM Deep Blue just before 2000
- IBM Watson in this millennium (from 2011)
29. Origin of IBM Watson
Source:
https://www.youtube.com/watch?v=JHahhixeQZs
30. • 3 guys play a match, based on 3 games, where they have to answer to some
difficult questions, where a human has to test his cross knowledge of arguments
•One of the 3 guys chooses a specific «category» of unknown questions, plus a
level of difficulty from 1 to 5
•The presenter makes the question, and who books first the possibility to talk
gives the answer, winning or losing the associated money based on the answer
•Of course the winner is the guy who earned more money
•The first development of Watson by IBM was done to be tested with Jeopardy,
so you saw Watson as one of the playing guys with a strongly enhanced human
behaviour, based on its cognitive engine
How to play to Jeopardy
31. celebrated
India
In May
1898
400th
anniversary
arrival in
Portugal
arrived in
India
In May
In May, Gary arrived in
India after he celebrated
his anniversary in
Portugal.
Garyexplorer
celebrated
anniversary
in Portugal
In May 1898 Portugal
celebrated the 400th
anniversary of this
explorer’s arrival in India.
Keyword Matching
Keyword Matching
Keyword Matching
Keyword Matching
Keyword Matching
Evidence suggests
that “Gary” could be
the answer.
BUT a good system
must also
understand that just
considering matching
words without
mathematics or
correlations is not
enough to obtain the
right answer!
Question Evidence
Hypotesis of answer based on facts and evidences in text
32. On the 27th of May 1498, Vasco da
Gama landed in Kappad Beach.
On the 27th of May 1498, Vasco da
Gama landed in Kappad Beach.
celebrated
India
May 1898 400th anniversary
arrival in
Portugal
landed in
27th May 1498
Vasco da Gama
Temporal
Reasoning
Statistical
Paraphrasing
GeoSpatial
Reasoning
explorer
Make the search more perfect
Extend hypotesis possibilities
Use different algorithms
Evaluate evidences
On the 27th of May 1498, Vasco da
Gama landed in Kappad Beach.
Kappad Beach
In May 1898 Portugal
celebrated the 400th
anniversary of this
explorer’s arrival in India.
Data math
Parapharase
GeoKB
Anyway, this is not yet 100% sure! Other facts and evidences might be found!
With the
support of
mathematics
and
correlations,
the answer gets
much better
taking to
“Vasco da
Gama”
Question Evidence!
Mathematics and correlations mandatory element to build a
cognitive system
33. Engine to generate and evaluate «answers hypotesis» combining more than 100 different alghoritms of language analysis,
information retrieval, machine learning and reasoning. Analytics resources collect, evaluate, estabilish weight and balance
different types of evidences in order to give the best possible answer where all information stored and elaborable by its cognitive
intelligence concurr to the result.
Answer
Scoring
Models
Answer &
Confidence
Question
Evidence
Sources
Models
Models
Models
Models
Models
Primary
Search
Candidate
Answer
Generation
Hypothesis
Generation
Hypothesis and
Evidence Scoring
Final Confidence
Merging &
Ranking
Synthesis
Answer
Sources
Question &
Topic
Analysis
Evidence
Retrieval
Deep
Evidence
Scoring
Learned Models
help combine and
weigh the Evidence
Hypothesis
Generation
Hypothesis and Evidence
Scoring
Question
Decomposition
1000’s of
Pieces of EvidenceMultiple
Interpretations
100,000’s Scores from
many Deep Analysis
Algorithms
100’s
sources
100’s
Possible
Answers
Balance & Combine
Technology of Watson – A Massively Parallel Probabilistic
Evidence-Based Architecture
34. Diagnosis supportDetecting Frauds
Support discovery
and analysis of
evidence
Analysis and correlation
of criminal investigation
facts and data
Customer Service Knowledge management
Several possible uses of IBM Watson Analytics
• Analysis, decomposition
and reconstruction of
events
• Generating Hypothesis
• Knowledge management
• Decision Support
• Deep Q & A
• Systems of questions and
answers ....
• On the right some
examples
36. You can easily try yourself IBM Watson Analytics
Create a free account for using IBM Watson
– http://www.ibm.com/analytics/watson-analytics/
Confirm your email ID to IBM Watson
Connect and watch the video
Start your test with sample data or your data
Connect with IBM Watson Analytics
–Facebook
–Twitter
–LinkedIn
38. Centralized,
Distributed, Cloud, Security, Social,
Resilient Architectures Increase Data
Volume
Determining problems in the third millennium
38
Where do I
start??
Everything is
“green”
It’s SLOW!!
404 ERROR
Determining problems
using analytics
blog:
http://ibm.co/1LBiiBQ
40. How to make problem determination faster with a machine
The IBM IT Operation Analytics Cognitive tool INGESTS big
data information coming from system logs
A google-like USER INTERFACE is available for search
- Ingested data
- Manuals
- Internet
The user interface transforms row data in readable information
- Statistics
- Sort and order information
- Filter to go directly to the cause of the problem
Several figures can use the tool
- Systems Engineer
- Subject Matter Expert
- Application expert
- Executive person
41. Search Workspace – Search, Navigate, Visualize
Search specific
logs or ALL logs
Timeframe
Enter search string
Save My Search
Simple search interface
EASY to customize
42. Raw data after a search through the analytic system
• The first search starts to simplify, but still too hard
• Need of a better way to analyze what the system has learnt
43. Switch to List View Column headers
derived from annotation
Search Filters
For easy
drilldown on
java
exceptions
Apps
Filtered and ordered results and insights