1) The document discusses reasons to invest in AI now, how companies can adopt AI, and the future of AI.
2) Key reasons to invest now include increased computing power, large datasets, and major companies investing billions in AI. AI could contribute $15 trillion to global GDP by 2030.
3) Companies should start with small automation projects to reduce costs and errors, then expand use of AI. They should also create cross-functional product teams to identify opportunities.
4) The future of AI includes more advanced voice recognition, predictive platforms, expanded use in industries like manufacturing and education, and applications in healthcare like disease diagnosis and drug development. AI is already used widely in applications like digital assistants,
1. Who & Why do I care ?
1
Hidden Figures
Marvin Laucher, part time coal miner, part time website builder.
C++
CNN
Convolution Neural
NetworkRef – Nasa.com, Cnn Money,youtube.com
3. Successful if we discuss
1. Why should I invest in AI now ?
2. How can I use AI research to help my company?
3. What should I expect in the future ?
3
Goal: Provide a perspective on AI adoption
6. AI is not new - Why now?
6
Ref – McKinsey Co, MIT Lex Fridman, HBR, Indeed.com
1. Computing scale: CPU, GPU,
ASICs
2. Datasets and infrastructure to
handle big data
3. Amazon, Google, FB, MSFT
investing in platforms
7. $37B market by 2025
7
* - 2012 – 2017 ; Ref: Venture beat,
$15 B* AI investments with $15Trillion impact on GDP by 2030
Images: Intershala
Start ups
~$8B (2012-2016)
AutoTech
BI Analytics
(Training data)
Healthcare
2 31
Big Corporations
~$6B (2013 - 2016)
Voice is the
newText
AI Platform
Cloud
1
2
Vision & Image
recognition
3
Ford invested $1B in Argo self-drivingAI tech
8. Talent in high demand & ROI still too early
Reference - McKinsey & co HBR Survey, Engagednet
8
• AI can boost your top and bottom line
30% of the users scaling AI solutions are achieving revenue increase
Google launched “Auto Ads” which automatically places and selects ads /
formats (revenue lift of 10 percent with revenue increases ranging from five to 15
percent)
• # of open jobs in AI / ML / Data Science – 150K
• Digital capabilities come before AI
Odds of generating profit from using AI are 50% higher for companies that have
strong experience in digitization.
10. Academia meets business with AI
10
Researchers
Training Data Prep
Machine Learning
Deep Learning
Reading papers
Implement results
Businesses
Customer Pain
Corporate Strategy
Growth
Key Results such as
Profit, Revenue,
Basket size
• Google’s Kaggle - open source competitions
• Product Managers collaborate on use cases
Examine the end to end operations of the company and look for areas that can be automated
11. 11
Start Small - Save Cost /
Automation with error
reduction
Try new solutions in few
channels to showcase
results
Expand to other channels
& problem sets
Resulting web of trees
Create a product organization with agile practices & transparency with key objectives
Where can I use AI?
1. “Use Deep learning to automate anything a human can do without thinking much in less than 1 second e.g. Identify cat pictures or smile /
frown
2. Use for Predictive analysis “ – Andrew Ng with caveats
12. Sample product purchase evolution
12
Door to door sales or
In-Store
Online & Mobile platform
Crowd sourced Marketplace
(Amazon, Ebay, Etsy,
Nextdoor, Facebook)
Predictive Marketplaces
Predictive
Voice
Alexa
Google
Home
iOT
Dash (Human
in the loop)
Vending
Machines
Robots
Inventory
Robots
Pepper
Robot
13. When should I stop the algorithm?
Ref – Andrew Ng
13Time
Human Level Performance
Human level accuracy
Optimal Error Rate / Bayes Error Rate
16. AI is already in our everyday lives
16
Health & WellnessBasics
• DNA sampling & diagnosis
• Lifestyle predictive Management
• ER & Hospital management
• Remote Doctors
• Mental Health
• Drug Discovery
• Shopping Recommendations
• FB image recognition
• Siri Voice recognition
• Google Home / Alexa assistants
• Optimized path maps
• Spam filter on email
• News aggregators / Personalized
news
17. 17
Next 5 and 10 years
• Voice recognition & ease of use
• Platforms to prepTraining data &
Analyze
• AppliedAI in Finance, Security,
Education (Block chain will aid this)
• Industry 4.0 (Manufacturing /
Weaponry…)
• Autonomous vehicles
• AR/VR in projection spaces (don’t need to
wear devices)
• Vision & Emotion precision: E-commerce
(Electronic malls), Intelligent individual
• Healthcare Disease diagnosis
• Distant Doctor / Drug development
• Disease free longevity (cancer cure)
• Space exploration
• Environment protection (if we live
on Earth)
5+ 10+
18. 18
Unite using technology to solve big problems without
dehumanizing
Ref – http://blog.crisman.com image with edits
21. HBR survey summary of 3000 AI executives
Reference - McKinsey & co
21
• Don’t believe the hype: Not every business is using AI despite a 26B investment in AI
20% are using at least 1 solution at scale
• Believe the hype that AI can potentially boost your top and bottom line
30% of the users are achieving revenue increase
• Without support from leadership, your AI transformation might not succeed
• You don’t have to go it alone on AI — partner for capability and capacity. Machine learning is a powerful tool, but it’s
not right for everything.
• Resist the temptation to put technology teams solely in charge of AI initiatives
• Take a portfolio approach to accelerate your AI journey
Short-term: Focus on use cases where there are proven technology solutions today and scale
Medium-term: Experiment with technology that’s emerging but still relatively immature
Long-term: Partner to solve a high-impact use case with bleeding-edge AI technology
• Digital capabilities come before AI
Odds of generating profit from using AI are 50% higher for companies that have strong experience in digitization.
24. Identify data needs & create a service mentality
Data Hygienists
Clean incoming data for accuracy
E.g. Calendar days vs. working days to count # of days
Data Explorers
Sift through data to discover the data we actually need
E.g.Training data
Solution
Architects
Organize the explored data for analysis & querying
Data Scientists Model the organized data for predictive analytics
Experience
Experts
Turn the models into experiences that get results
E.g. e-mail, Interactive
Reference – HBR
24
25. Auto-tech, BI & Healthcare are
major investment areas
25
Investments By Sectors in USD MM
27. Robotics, Auto, Facial & Voice
recognition attract huge investments
27
Investments Size in USD MM
Investments Size in USD MM
Number of companies by investment size
29. Voice will be the new Text
29
Google
• 11+ acquisitions
• ML Platform creation
• Vision / Image and Speech recognition
• Business Process improvements
Apple
• 7+ acquisitions
• Vision / Image and Speech
recognition
• Catch up with Google on
platform creation (Turi)
Facebook
• Vision / Image and Speech recognition
• Voice activation SDKs
Microsoft
• Voice enabled assistant
• Type ahead predictor
• Voice activation SDKs (AI
Fund)
30. 30
These companies market cap surpass the GDP of
India (previously Russia and Canada)
Reference – Scott Galloway Ted talk
Dorothy Vaugh
Hidden figures Fortran
Coal Miners Computer
C++ CNN
Corporate Social Responsibility to learn and let others learn.
Google gets 1 new question for every 6 questions – Scott Galloway. 1.3 to 1.4 Billion per second
Google and Facebook have made ¼ of India’s GDP in market cap due to our love of these products
3.5 B searches per day
Now we have the 700,000 best and brightest, and these are the best and brightest from the four corners of the earth. They are literally playing with lasers relative to slingshots, relative to the squirt gun. They have the GDP of India to work at. And after studying these companies for 10 years, I know what their mission is. Is it to organize the world's information? Is it to connect us? Is it to create greater comity of man? It isn't. I know why we have brought together -- I know that the greatest collection of IQ capital and creativity, that their sole mission is: to sell another fucking Nissan.
Large data set & processing tools
Modern algorithms: Backprop, CNN, LSTM
Infrastructure / Software
With 15 Trillion impact by 2030 and 15 B in Investments (8 B in startups and 6 – 7B in large companies), we can determine how the funding is distributed.
Top 3 areas will be:
Auto tech – self-driving cars
Training data prep and prediction on open source data using block chain for security (Ayushnet)
Healthcare – Early prevention and diagnosis
The adoption is slow because of the
digital proficiency (legacy platforms)
Non unified data sources
learning curve and talent
Platforms are not standardized
Leadership support missing
Gartner Says More Than 40 Percent of Data Science Tasks Will Be Automated by 2020
Approximately 32058KBy 2018 alone we could be short of about 140K to 190K people with deep analytical skills according to McKinsey Global Institute
Don’t believe the hype: Not every business is using AI despite a 26B investment in AI
20% are using at least 1 solution at scale
Without support from leadership, your AI transformation might not succeed
You don’t have to go it alone on AI — partner for capability and capacity. Machine learning is a powerful tool, but it’s not right for everything.
Resist the temptation to put technology teams solely in charge of AI initiatives
Andrew Ng
Use Deep learning to automate anything a human can do without much thinking less than 1 second e.g. Identify cat pictures or colors
Predictive analysis
Retailers use both physical automation such as robots in warehouses and algorithms to predict what users will purchase
Where can you automate by augmenting human wisdom?
Big data initiatives fail because the internal customers don’t have confidence in the team and don’t trust the models. Trust starts with transparency - open about who is working on what.
Align incentives of the Data team with business teams. Are your models being used by business?
WORRY LESS ABOUT CRUNCHING IT BUT MORE FOCUSSED ON SERVING YOUR MODEL.
In statistical classification, Bayes error rate is the lowest possible error rate for any classifier of a random outcome (into, for example, one of two categories) and is analogous to the irreducible error. A number of approaches to the estimation of theBayes error rate exist. = Optimal Error Rate
Theoretical limit of performance e.g. blurry images or sound that is not recognizable and that level is call optimal error rate (the line above the human error rate)
Intelligent Individuals – The computer is analyzing sentiment, emotion and recommending right action or learning function (read ABC)
Unite to solve
Where are the opportunities?
What are the types of data roles?
Big data initiatives fail because the internal customers don’t have confidence in the team and don’t trust the models. Trust starts with transparency - open about who is working on what.
Align incentives of the Data team with business teams. Are your models being used by business?
WORRY LESS ABOUT CRUNCHING IT BUT MORE FOCUSSED ON SERVING YOUR MODEL.
BI & Analytics – Training data preparation and prediction on open source data
Top 3 areas will be:
Auto tech – self-driving cards
Training data prep and prediction on open source data using block chain for security (Ayushnet)
Healthcare
Drill down to healthcare
- Disease diagnosis
Predictive Analytics
Remote monitoring and data analysis
Drug development
1. Predicted that there are more number of startups in early stages and some move on to multiple series or just get huge investments because they require heavy Capex like auto / robotics or Industry 4.0 manufacturing robots.
2. I wish there were more investment in Education consumer and health tech AI products which help everyday people reduce the education gap / health gap.
3. Where do you want to see the future AI
- Space exploration
Brain is Google’s automated algorithm learning system. RankBrain is used for search queries
Business process improvements with Deepmind acquisition e.g. energy savings with data
Vision / Speech / Image recognition AI acquisitions
Google
11+ acquisitions
ML Platform creation
Vision / Image and Speech recognition
Business Process improvements
Apple
7+ acquisitions
Vision / Image and Speech recognition
Catch up with Google on platform creation (Turi)
Facebook
Vision / Image and Speech recognition
Voice activation SDKs
Microsoft
Voice enabled assistant
Type ahead predictor
Voice activation SDKs (AI Fund)