earley executive roundtable on data analytics - session 1 - the business potential of machine...

Copyright © 2015 Earley Information Science1 Copyright © 2015 Earley Information Science

Earley Executive Roundtable

Series on Data Analytics

Session 1: Business Potential of Machine

Learning and Cognitive Computing

May 27, 2015

Presented by

Seth Earley

CEOClick to watch the

recording of this session

http://info.earley.com/earley-roundtable-session-1-business-potential-of-machine-learning-cognitive-computing

Copyright © 2015 Earley Information Science2

Today’s Agenda

• Welcome & Housekeeping

– Session duration & questions

– Session recording & materials

– Take the survey!

• Introduction – Seth Earley

• Panelist Introductions

– Bruce Daley Principal Analyst, Tractica (@brucedaley)

– Olly Downs, Chief Scientist/CTO, Globys (@globysinc )

– Mitchell Shuster, Data Scientist, Knowledgent (@Knowledgent)

– Patrick Heffernan, Practice Manager, TBR (@TBR_PatrickH)

• Panel Discussion

• Questions & Answers


Seth Earley, Founder & CEO, Earley Information Science

[email protected]@sethearley

• Over 20 years experience in data science and technology, content and knowledge

management systems, background in sciences (chemistry)

• Current work in cognitive computing, knowledge and data management systems,

taxonomy, ontology and metadata governance strategies

• Co-author of Practical Knowledge Management from IBM Press

• Editor of Data Analytics Department IEEE IT Professional Magazine

• Member of Editorial Board Journal of Applied Marketing Analytics

• Former Co-Chair, Academy of Motion Picture Arts and Sciences, Science and

Technology Council Metadata Project Committee

• Founder of the Boston Knowledge Management Forum

• Former adjunct professor at Northeastern University

• Guest speaker for US Strategic Command briefing on knowledge networks

• AIIM Master Trainer – Information Organization and Access

• Course Developer and Master Instructor for Enterprise IA and Semantic Search

• Long history of industry education and research in emerging fields


Machine Learning and Cognitive Computing

Core Concepts


Machine Learning - The detection of patterns and surfacing of information through a

variety of approaches based on statistics and mathematics

A search index is a derivation of

structure from unstructured

information (clustering, classification,

entity extraction and various text

analytics approaches use machine

learning approaches)

Advanced search algorithms detect

“signals” from users’ intent and past

search patterns to increase the

relevance of search results


Machine Learning - The detection of patterns and surfacing of information through a

variety of approaches based on statistics and mathematics

“More like this” and “users who liked

this also liked that” types of results

leverage machine learning algorithms

Systems that classify documents

based on “training sets” use analytical

methods to create mathematical

representations of content and

documents

Personalization – content, search

results or product recommendations

are all based on a system for

“predicting” what you are looking for.



Siri answers your questions about movie

times, sports scores, restaurants nearby

Cognitive Computing - A way for computers to be more user friendly and “understand”

what humans want

Watson answered tricky and ambiguous

trivia questions with obscure references,

puns, metaphors, time references, slang,

idiomatic expressions and other

challenging types of ambiguous queries

Interpreting signals - what a user is

looking for in a query, interpreting

questions asked in plain English (natural

language), engaging in a dialog,

“understanding” the meaning of an

ambiguous question, anticipating the next

step in a process

Pattern recognition,

pattern matching and

rules for predicting

outcomes



Artificial intelligence

encompasses all of these tools

and techniques to solve various

types of problems – from

writing articles to driving cars to

detecting fraud, diagnosing

disease, making decisions that

have previously been in the

realm of human judgement


Today’s Panel of Experts

Bruce Daley, Olly Downs, Mitchell Shuster, Patrick Heffernan


Bruce Daley

• Contributor to Tractica’s Automation & Robotics practice with

focus on artificial intelligence and machine learning for enterprise

applications

• Previously, vice president and principal analyst with Constellation

Research covering business research themes related to

customer relationship management, mobility, and infrastructure

• Also, founder of Great Divide, co-founder of Rabbit Ears Capital

Advisors, founder of Test Common Inc., founder of the Enterprise

Software Summit, and founder of The Siebel Observer, the

largest publication devoted to Siebel Systems

• Additionally, held consulting and management roles at Oracle

and Bain & Company

• Widely quoted industry expert in major publications including The

Wall Street Journal, The New York Times, The Financial Times,

The International Herald Tribune, IEEE Spectrum, The San Jose

Mercury News, and many more.

• Author of a soon-to-be-published book on data storage, Where

Knowledge is Power, Data is Wealth

• Holds a BA from Tufts University

Principal Analyst

Tractica

@brucedaley


Ad Services Automotive Agriculture Finance Data Storage

Education Investment Health Care Legal Manufacturing

Media Medical Oil and Gas Philanthropy Retail

• Self driving cars• Self parking cars• Diagnostics

• Control plane• Watson

• Credit scoring• Fraud detection

• Personalized geo location

• Intelligent agents

• Forecasting• Fraud detection

• Rotor position estimation

• Predicting electricity prices

• Optimizing milling parameters

• Testing mathematical theorems

• Grading exams• AI tutors• Gamification

• Writing sports stories

• Storytelling analysis

• Analyzing seismic data

• Estimating size oil reservoirs

• Determine min gas miscibility

• Fundraising• Optimized giving• Smart charity• Moral AI• Fraud detection

• Malignant pleural mesothelioma

• Orthodontic diagnosis

• Lung CT classify

• Clinical trial compliance

• Adverse drug reaction prediction

• Crop planting optimization

• Develop drought tolerant crops

• Self driving tractors

• Electronic discovery

• Patent infringement analysis

• Contract

• Program trading• High frequency

trading• Algorithm trading• Index arbitrage

• Personalized ad serving

• Ad portfolio optimization

POV – Bruce Daley

My point of view – after years of false starts, the amalgamation of statistics, GPU chips, deep

learning algorithms, and big data has made narrow applications of AI practical. The only limit to

the problems they are being asked to solve to seems to be the human imagination

New Algorithms

GPUDataDEEP

LEARNING

Probability and Statistics


Dr. Olly Downs

• Responsible for the analytics strategy, technical approach and

algorithm design and development for Globys’ marketing

personalization technology platform (Amplero).

• A machine learning scientist and serial technology entrepreneur,

credited with bringing advanced analytics and machine learning

methods to bear as the creative spark behind numerous early-stage

technology companies.

• Specializes in applying abstract analytical ideas from mathematical,

physical and statistical science to problems in the real world and

commercializing them into significant businesses.

• Recently served as Chief Scientist at Atigeo, Chief Scientist at

Mindset Media (sold to Meebo, February 2011) and Director of

Research at Pelago (sold to GroupOn, April 2011).

• As Principal Scientist at INRIX, the first technology spin-out from

Microsoft Research, delivered a world-first in the provision of real-

time traffic information using a nationwide network of GPS-enabled

probe vehicles.

• Holds Ph.D. and MA degrees in Applied & Computational

Mathematics from Princeton University, and BA, MA and MSci

degrees in Experimental & Theoretical Physics from the University

of Cambridge, UK.

Chief Scientist/CTO

Globys

@globysinc

http://www.amplero.com/


Olly Downs – POV

• Successful adoption of Machine Learning (ML) and Cognitive Science to

drive business value is split across 3 segments

– Large businesses for which these capabilities are core to their business

– Businesses for which these capabilities are strategic and that can invest in team and

tools

– Businesses for which these capabilities are valuable but inaccessible

• “Chasm” between 1 & {2,3}

– Getting business-impacting results and operationalizing is difficult

– Processes are unwieldy, and even best practices with teams and tools move slowly

i.e., weeks, months

– Hiring and retaining people with the right skills is not easy (as #1 consumes them)

• Proposition for ML and Cognitive Computing to “Cross the Chasm”

– Add value without hands on intervention – discover and act without human experts

– Inform and educate on what is discovered

– Reduce upfront investment hurdle

[ 13 ]


Example – Applying Machine Learning to

Marketing

[ 14 ]

Plan Development LaunchTraditionalCampaign

Process

TestOptimize Design

TestAnalyze

Analyst/Data Scientist

Plan DevelopmentLaunch

Campaign Process

With Amplero

Discover

“Team & Tools” Approach:Weeks of Design and Configuration10’s of Marketing ContextsWeekly AnalysisOne-time OptimizationBI Team Researching Results

Machine Learning Approach:Configuration in Minutes1,000’s of Marketing ContextsDaily DiscoveryContinuous OptimizationBI Team Researching New Revenue Opportunities


Mitchell Shuster

• Award-winning data scientist, physicist, and technology

entrepreneur who seamlessly blends analytic and research

knowledge honed in the academic realm with real-world technical

and industry expertise.

• As Informationist and Data Scientist at Knowledgent, the data and

analytics firm, specializes in applying advanced analytics and data

science concepts and techniques, including machine learning

(Regression, Neural Nets, SVMs, Clustering, PCA, Anomaly

Detection, etc.), to help client organizations gain actionable insights

and competitive advantage.

• Currently leveraging predictive analytics expertise to deliver data-

driven models that improve patient outcomes, decrease costs, and

increase operational efficiency for healthcare and life sciences

organizations.

• Previously in Research & Development at Intel Corporation,

designed and developed basis for Intel's worldwide high-volume

manufacturing at the newest technology node and was recognized

for computational modeling and process implementation.

• Earned Ph.D. degree in Physics and multiple research fellowships

from Penn State University, where he authored research published

in multiple prominent peer-reviewed scientific journals, and BA

degree in Physics from Cornell University.

Informationist and

Data Scientist

Knowledgent

@Knowledgent


POV – Mitchell Shuster

• Machine learning is a powerful tool for extracting

information from data

– It is non-trivial to frame the questions, prepare the data, and interpret

the result in context

– The right data is required to answer a given question

– “Machine learning” is not a magic wand to solve all problems

• Cognitive computing is an extension of compute

capabilities into more human-like interactions

– The primary distinguishing characteristics are context awareness

and tolerance of ambiguity

– At present, limited to specific tasks and contexts

Beware the hype! What is possible is not always practical.

What is practical is not always desirable.


Patrick Heffernan

• Coverage and Focus Areas include IT Services, management

consulting, global delivery, strategy and operations, cloud,

intelligence cycle, project management, and client engagement

• Directs the practice’s syndicated portfolio and cultivates and

manages projects on topics ranging from management consulting

to firms’ financial advisory services to emerging technologies.

• Expertise in competitive intelligence, strategy, and global political-

economic impacts on business cycles and consulting vendors.

• Prior to joining TBR, was part of a Big Four firm’s competitive

intelligence team, conducting field work and analysis.

• Professional career started in diplomacy, with Middle East postings

as a foreign service officer with the State Department and

counterterrorism assignments with the National Security Council

and the U.S. Department of the Treasury.

• Received a B.A. from Washington and Lee University and an M.A.

in foreign affairs from the University of Virginia.

Practice Manager and

Principal Analyst,

Professional Services

Practice

Technology Business

Research

@TBR_PatrickH


POV – Patrick Heffernan

• Cognitive computing and machine learning will

increasingly have business impacts on ---

– IT services vendors, including Accenture, Infosys, Wipro, IBM, and

Cognizant, as those vendors must invest in people and capabilities

to keep pace with competition and grow in new areas -- and these

vendors are afraid of being too slow, too late, or too “me-too” for the

market;

– clients who appreciate the potential of what the vendors listed above

can deliver, but don’t know how disruptive those changes will be –

these companies are afraid of being too aggressive in adopting

emerging technologies and paying a premium for what will soon be a

commodity; and

– employees at IT services vendors and at their clients who fear losing

their jobs to “robots” – this is a recurring fear when emerging

technologies take root, but just because it keeps coming up doesn’t

mean it isn’t real


Discussion

• OK, interesting stuff - where do I get started?

• How do I tell what is possible from what is practical and achievable for

my organization?

• What kinds of problems can I solve?

• What is the difference between “deep learning” and “machine learning”?

• What kinds of education does my team need? Where do I get it?

• What are the industries and applications that are most mature?


Thank you to our sponsors/producers

www.computer.org/itpro

http://www.henrystewartpublications.com/ama

www.informationdevelopmentworld.com

www.thecontentwrangler.com

http://www.tbri.com

http://www.computer.org/itpro

http://www.henrystewartpublications.com/ama

http://www.informationdevelopmentworld.com/

http://www.thecontentwrangler.com/

http://www.tbri.com/


For more information

• IT Professional Magazine - www.computer.org/itpro Next issue focuses on Analytics

• Computing Edge http://www.computer.org/web/computingedge (highlights of IEEE

publications)

• Cognitive Computing and Big Data Analytics by Judith Hurwitz, et al

http://www.amazon.com/Cognitive-Computing-Big-Data-Analytics/dp/1118896629

• Artificial Intelligence for Enterprise Applications https://www.tractica.com/research/artificial-

intelligence-for-enterprise-applications/ (contact [email protected] mention roundtable to

get 10% discount)

• Microsoft- Machine learning blog:

http://blogs.technet.com/b/machinelearning/archive/2015/05.aspx

• McKinsey- AI for the C Suite

http://www.mckinsey.com/insights/strategy/artificial_intelligence_meets_the_c-suite

• Stanford course in machine learning https://www.coursera.org/course/ml

• Data science and machine learning resources: http://conductrics.com/data-science-

resources/

• Video lectures: http://videolectures.net/Top/Computer_Science/Machine_Learning/

http://www.computer.org/itpro

http://www.computer.org/web/computingedge

http://www.amazon.com/Cognitive-Computing-Big-Data-Analytics/dp/1118896629

https://www.tractica.com/research/artificial-intelligence-for-enterprise-applications/

mailto:[email protected]

http://blogs.technet.com/b/machinelearning/archive/2015/05.aspx

http://www.mckinsey.com/insights/strategy/artificial_intelligence_meets_the_c-suite

https://www.coursera.org/course/ml

http://conductrics.com/data-science-resources/

http://videolectures.net/Top/Computer_Science/Machine_Learning/


Mining Business Insights with Big Data Analytics and the

Internet of Things

Joanna SchlossBusiness Intelligence and Analytics Evangelist, Dell Software

John SpoonerVice President, Platforms, Technology Business Research, Inc.

Ram SangireddyDir of Product Management, Predictive & Analytics, Vitria Technology

Bruce DaleyPrincipal Analyst, Tractica

Next Session: June 3rd 1pm EDT


Earley Information Science helps

organizations establish a strong

information architecture and

content management foundation

Specializing in making information more findable,

useable and valuable to drive digital commerce

innovation, enhance customer experience, and

improve operational efficiency and effectiveness.

Realize your digital transformation vision

with EIS.

Earley Information Science

(EIS)A trusted information integrator

Founded – 1994

Headquarters – Boston, MA

www.earley.com

Seth Earley, CEO

Email: [email protected]

Twitter: @sethearley

LinkedIn: www.linkedin.com/in/sethearley

mailto:[email protected]

http://www.linkedin.com/in/sethearley


A Broad Spectrum of Business Solutions

DIGITAL BUSINESS SOLUTIONS

B2C Digital Commerce

• Product Curation for a World-Class Product Catalog

• Site Merchandising Taxonomy & Attribute Design

• Information Architecture for Shopper Context

B2B Digital Commerce

• Product Search & Findability

• Product Information Management

• Product Knowledge Management

Digital Workplace

• Enterprise Content & Records Management

• Information Architecture

• Enterprise Knowledge Management


EIS Reference Architecture

earley executive roundtable on data analytics - session 1 - the business potential of machine...

Data & Analytics