Date post: | 21-Mar-2017 |
Category: |
Technology |
Upload: | adrienne-debigare |
View: | 1,731 times |
Download: | 4 times |
Exploring opportunities in cognitive technologies
November 2015
Agenda
What is AI? What are cognitive technologies?
Focus on machine learning The catalysts of progress Types of applications Investment and trends Where and whether to apply these technologies Cog tech, automation, and work Conclusion
What is artificial intelligence?DEFINING AI
“… may lack an agreed-upon definition” − AI pioneer Nils Nilsson1
Leading AI textbook offers 8 definitions2
1 Nils Nilsson, The Quest for Artificial Intelligence2 Stuart Russell and Peter Norvig, Artificial Intelligence
A useful definition of AIDEFINING AI
The theory and development of computer systems able to perform tasks that normally require human intelligence.
The AI EffectDEFINING AI
“As soon as it works, no one calls it AI anymore.”1
“AI is whatever hasn’t been done yet.”2
1 John McCarthy, quoted in Bostrom2 Attributed to Larry Tesler in Douglas R. Hofstadter, Godel, Escher, Bach: An Eternal Golden Braid
Cognitive technologies
Cognitive technologies simulate perceptual and cognitive skills to perform tasks only humans used to be able to do
Cognitive technologies
Such as
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Supervised learning
Supervised learning is like learning by example“Learning a model from labeled training data”
Used for• Classification - output is
one of set of discrete values (e.g. spam, not spam)
• Regression - output is a number (e.g., a price) - prediction
Source: http://faculty.chicagobooth.edu/drew.creal/teaching/basiccoursematerial/lectures/lecture9.pdf
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Unsupervised learning
Learning by discovering patterns – “There are two types of people ….”
Applications - customer segmentation, product basket discovery, topic analysis
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Reinforcement learning
Learning by trial and error – how a baby learns to crawl
Rights:https://pixabay.com/en/baby-crawling-child-infant-shadow-147416/
Applications:- Mechanical control - elevators, robots- Game playing
Catalysts of progress in AI
Moore’s law benefited all types of computingCATALYSTS OF PROGRESS
Current generation of microprocessors are 4,000,000X more powerful than first single-chip microprocessor of 19711
1 Andrew Danowitz et al., “CPU DB: Recording microprocessor history,” ACMQueue, volume 10, issue 4 (April 6, 2014), http://queue.acm.org/detail.cfm?id=2181798, accessed October 11, 2014.
Big data and new techniques advance work in AICATALYSTS OF PROGRESS
• Volume of data doubles every 2 years1
• 44 trillion gigabytes annually by 20202
• New techniques for managing and analyzing data
• AI models improved with “training”
1, 2 IDC 2014, http://www.emc.com/leadership/digital-universe/2014iview/executive-summary.htm
Internet and cloud support AI with access to big data and collaborators
CATALYSTS OF PROGRESS
Access to vast data resourcesCrowd-sourcing to train machine learning modelsImplicit collaboration, e.g., Web search, translation
Advances in algorithms broke performance barriers
CATALYSTS OF PROGRESS
New algorithms dramatically improve performance of machine learning
Over 500,000 scholarly papers on neural networks since 20061
New distributed computing breakthroughs
1 Google Scholar
Performance is improving…continually
Facial recognition: 2014: 97% accuracy (Facebook)1; 2015: 100% accuracy (Google)2
Google speech recognition:2013: 23% error rate2015: 8% error rate3
IBM Watson2400% smarter than when it won Jeopardy!4
1 Facebook, “DeepFace: Closing the gap to human-level performance in face verification,” https://www.facebook.com/publications/546316888800776/, accessed October 3, 2014; 2 http://www.cs.toronto.edu/~ranzato/publications/taigman_cvpr14.pdf; 3 Jordan Novet, Venture Beat, “Google says its speech recognition technology now has only an 8% word error rate,” May 28, 2015, http://venturebeat.com/2015/05/28/google-says-its-speech-recognition-technology-now-has-only-an-8-word-error-rate/, accessed September 16, 2015; 4 IBM, “IBM Watson,” http://www-03.ibm.com/press/us/en/presskit/27297.wss, accessed October 3, 2014.
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Three main applications of cognitive technologies: Product, process and insight
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Product: Embedding cognitive technologies in a product or service
Embed cognitive technologies to help increase the value of products or services by making them more effective, convenient, safer, faster, distinctive, or otherwise more valuable.
eBayNetflix
GM Domino’s Pizza
AudiVuCOMP
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Process: automate tasks or processes humans used to do
Using computer systems to do work that people used to do. The work gets done faster, cheaper, better, or some combination of the three. Organization benefits.
Automate scheduling engineering works
Clinical trials eligibility
Process handwritten forms
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Insight: Discerning patterns, making predictions, to improve operations or guide strategy
Generate insights that can help reduce costs, improve efficiency, increase revenues, improve effectiveness, or enhance customer service• Intel• BBVA Compass• Aetna
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
New applications of machine learning daily
Public sector• Smartphone app to reduce urban congestion• Flag parking abuses• Detect misbehavior by prisonersAutomotive• Detected driver absent-mindednessEnterprise information management• Classify and route business documents• Setting rules for accessing and manipulating
documentsHealth• Detect signs of gambling addiction• Predict cancer remission or drug resistance• Drug discovery• Predict development of psychosis• Predict air pollution days in advanceSales• Predicting which deals will close
Announced in the last few months
Billions in investment capital aimed mostly at traditional business problems and industries
Supporting Technologies
Rethinking Humans / HCI
Core Technologies
Rethinking Industries
Rethinking Enterprise
$281,293,137.0
$855,050,003.0
$1,037,431,892.0
$2,000,530,027.0
$2,464,000,488.0
VC investment in cognitive technology companies that have raised at least $10M, (Jan. 2011 – Sep. 2015, US only)
Commercialization and improving performance expand applications
Improving performance and commercialization fueled by surging investment expand the applications for cognitive technologies
GRAPHIC: DELOITTE UNIVERSITY PRESS
Applications are broadening
As performance improves, applications of speech recognition, computer vision, natural language processing and understanding are growing
conclusion
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Deciding where to apply cognitive technologies in an organization
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Viable: Where is it possible to apply cognitive technologies
Types of tasks ExamplesPerceptual tasks involving vision, speech, reading handwriting
Forms processing, first-tier customer service, warehouse operation
Analytical tasks, involving large data sets
Document review; finding patterns, making predictions
Decision-making tasks where expertise can be expressed as rules
Planning maintenance operations
Planning tasks in a constrained domain
Scheduling
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Valuable: Where it is worth applying cognitive technologies
It may be worth using cognitive technologies where• Workers’ cognitive abilities or training are
underutilized• Business process has high labor costs• The value of improved performance is high
Opportunities ExamplesWorker’s cognitive abilities or training are under utilized
Writing company earnings reports; e-discovery
Business process has high labor costs
Medical utilization management
Expertise is scarce Medical diagnosis—especially rare conditions
Value of improved performance high
Decision-making in financial services
Create new features customers care about
Natural interfaces, automation, “intelligence”
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Consumer benefits of cognitive technologies
Anupam Narula, David Schatsky, Ben Stiller, & Robert Libbey, "The thinker and the shopper: Four ways cognitive technologies can add value to consumer products," Deloitte University Press (June 3, 2015)
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Vital: Where it is necessary to apply cognitive technologies
It may be necessary to use cognitive technologies where• Industry-standard performance requires it
(e.g. Online retail product recommendations)• Cannot scale relying on human labor alone
(e.g. media sentiment analysis, fraud detection)
Types of tasks ExamplesIndustry-standard performance requires cognitive tech
Online retail product recommendations
Service cannot scale relying on human labor alone
Fraud detection; social media sentiment analysis
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
The unintended consequences of automation http://en.wikipedia.org/wiki/Automatic_control#/media/File:Control_room_pt_tupper.jpg
People are flawed; automated systems can have flaws too
Humans are bad at monitoring automated processes—paying attention to things that hardly change
People lose skills if they don’t practice them—the autopilot irony
Cognitive “underload” can reduce performance
Automated systems can undermine worker motivation, cause alienation, and reduce satisfaction, productivity, and innovation
Ill-conceived automation strategies may diminish our sense of self-worth
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
What to automate, and to what extent?
High 10. The computer decides everything, acts autonomously, ignoring the human,
9. informs the human only if it, the computer, decides to
8. informs the human only if asked, or
7. executes automatically, then necessarily informs the human, and
6. allows the human a restricted time to veto before automatic execution, or
5. executes that suggestion if the human approves, or
4. suggests one alternative
3. narrows the selection down to a few, or
2. the computer offers a complete set of decision/action alternatives, or
Low 1. the computer offers no assistance: humans take all decisions and actions
Information acquisition
Information analysis
Decision and action selection
Action implementation
Adapted from: Raja Parasuraman et al., “A Model for Types and Levels of Human Interaction with Automation,” IEEE Transactions on Systems, Man, and Cybernetics 30, no. 3 (2000): 286–297.
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Organizations have automation choices
Automation approach What is automated Examples
Replace Everything ATM; first-tier customer support
Atomize/automate As much as possible Machine translation plus human cleanup
Relieve Dull, dirty, or dangerous jobs
Routine earnings stories at AP; caller authentication at Barclays
Empower What wasn’t even being done before
IBM Watson for Oncology; oil & gas drilling problem resolution
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Maximizing the value of workers and machines
Besides automation choices, organizations must choose between a cost strategy and a value strategy
A cost strategy uses technology to reduce costs, especially by reducing labor
A value strategy aims to increase value by complementing labor with technology or reassigning labor to higher-value work
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Automation choices under different strategiesNEITHER THE TASK NOR THE TECHNOLOGY DICTATE THE STRATEGY TO BE FOLLOWED
Automation Choice
Cost strategy Value strategy
Replace Eliminate worker Reassign worker; use tech to provide superior performance
Atomize/ automate
Accelerate work, reduce staff, possibly alienate creative workers and artisans
Create new low-cost offers, employ lower-skilled, less-experienced workers
Relieve Eliminate routine tasks, increase productivity, reduce staff
Redeploy people to higher-value tasks; create more value for customers
Empower Increase performance of workers Increase workers’ performance and use to enhance their skills
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Some skills will become more valuableTAKING A FRESH LOOK AT WHAT SKILLS WILL BE NEEDED
TASKS THAT CANNOT BE SUBSTITUTED BY COMPUTERIZATION ARE GENERALLY COMPLEMENTED BY IT.
TECHNOLOGY INCREASES PRODUCTIVITY, RAISES EARNINGS, AND AUGMENTS DEMAND FOR SKILLED LABOR
THE SKILLS REQUIRED FOR ROUTINE WORK TO BECOME LESS VALUABLE
THE SKILLS REQUIRED TO PERFORM BROADLY-DEFINED JOBS AND THOSE REQUIRED FOR SUCCESSFUL INTERPERSONAL
INTERACTIONS TO BECOME RELATIVELY MORE VALUABLE:
Flexibility General problem solving
Creativity Tolerance of ambiguity
Empathy Drive
Emotional intelligence Resourcefulness
Critical thinking Openness to serendipity
Demystifying artificial intelligence: The real opportunities for cognitive technologies in business
Organizations must plan for cognitive technologiesTHERE ARE CHOICES TO BE MADE
COGNITIVE TECHNOLOGIES WILL CHANGE THE EMPLOYMENT LANDSCAPE IN THE COMING YEARS
SOME JOBS WILL DISAPPEAR; OTHERS WILL BE REDESIGNED; NEW KINDS OF WORK WILL ARISE
WORKERS WHOSE SKILLS ARE COMPLEMENTED BY COGNITIVE TECHNOLOGIES WILL THRIVE; THOSE WHOSE SKILLS ARE BEING SUPPLANTED BY SMART MACHINES MAY STRUGGLE
LEADERS FACE CHOICES ABOUT HOW TO APPLY COGNITIVE TECHNOLOGIES:
• WILL THEIR WORKERS BE MARGINALIZED OR EMPOWERED?
• WILL THEIR ORGANIZATIONS USE THE TECHNOLOGY TO CREATE VALUE OR CUTTING COSTS?
TALENT STRATEGIES MUST START TO ACCOUNT FOR IMPACT OF COGNITIVE TECHNOLOGIES
Some take aways
Understand how these technologies enable new, better ways of working.Prepare to adopt when appropriate, or risk being sidelined.
Begin today to explore cognitive technologies
But killer robots are not around the cornerSomething new and important is happening
Their impact on business is increasingThe technologies are getting better
An opportunity to differentiateThe use of cognitive technologies can confer competitive advantage today. It will become table stakes tomorrow.
New online course on artificial intelligence and cognitive technologies
Free course. Register today:
http://novoed.com/cognitive-technology
Now available on:
Signals for Strategists: Sensing Emerging Trends in Business and Technology This book is for strategists—leaders, managers, entrepreneurs—who are so caught up in the daily pressures of the business that they’re missing key signals of their future reality. Signals for Strategists identifies the emerging trends on the horizon. The sooner we see them, the better our response.
As used in this document, “Deloitte” means Deloitte Consulting LLP, a subsidiary of Deloitte LLP. Please see www.deloitte.com/us/about for a detailed description of the legal structure of Deloitte LLP and its subsidiaries. Certain services may not be available to attest clients under the rules and regulations of public accounting.
[email protected]: dschatsky
Copyright © 2015 Deloitte Development LLC. All rights reserved.Member of Deloitte Touche Tohmatsu Limited.
This publication contains general information only and Deloitte is not, by means of this publication, rendering accounting, business, financial, investment, legal, tax, or other professional advice or services. This publication is not a substitute for such professional advice or services, nor should it be used as a basis for any decision or action that may affect your business. Before making any decision or taking any action that may affect your business, you should consult a qualified professional advisor.Deloitte shall not be responsible for any loss sustained by any person who relies on this publication.