Starting from examples of current use cases of AI in business and in everyday life, we'll see what the future holds and we'll mention questions to address when giving autonomy to intelligent machines. We'll also aim at demystifying how AI works, in particular how machines can use data to automatically learn business rules and actions to perform in different contexts.

1. Future of AI-powered automation
in business
@louisdorard
#APIdays - December 9, 2015

2. AI is everywhere

10. @louisdorard

14. Lars Trieloff
@trieloff
(see source)

15. Amazon for David Jones (@d_jones, see source)

16. Amazon for David Jones (@d_jones, see source)

17. ChurnSpotter.io

18. How does it work?

19. Data + Machine Learning

21. Bedrooms Bathrooms Surface (foot²) Year built Type Price ($)
3 1 860 1950 house 565,000
3 1 1012 1951 house
2 1.5 968 1976 townhouse 447,000
4 1315 1950 house 648,000
3 2 1599 1964 house
3 2 987 1951 townhouse 790,000
1 1 530 2007 condo 122,000
4 2 1574 1964 house 835,000
4 2001 house 855,000
3 2.5 1472 2005 house
4 3.5 1714 2005 townhouse
2 2 1113 1999 condo
1 769 1999 condo 315,000

23. Bedrooms Bathrooms Surface (foot²) Year built Type Price ($)
3 1 860 1950 house 565,000
3 1 1012 1951 house
2 1.5 968 1976 townhouse 447,000
4 1315 1950 house 648,000
3 2 1599 1964 house
3 2 987 1951 townhouse 790,000
1 1 530 2007 condo 122,000
4 2 1574 1964 house 835,000
4 2001 house 855,000
3 2.5 1472 2005 house
4 3.5 1714 2005 townhouse
2 2 1113 1999 condo
1 769 1999 condo 315,000

24. ML is a set of AI techniques where
“intelligence” is built by referring to
examples

26. “Weak AI” vs. “Strong AI”

27. 27
Everyday use cases
• Real-estate
• Spam
• Priority inbox
• Crowd prediction
property price
email spam indicator
email importance indicator
location & context #people
Zillow
Gmail
Gmail
Tranquilien

28. 28
Business use cases
• Reduce churn
• Cross-sell
• Optimize pricing
• Predict demand
customer churn indicator
customer & product purchase indicator
product & price #sales
context demand
RULES

29. –Katherine Barr, Partner at VC-firm MDV
"Pairing human workers with
machine learning and automation
will transform knowledge work
and unleash new levels of human
productivity and creativity."

30. Decisions from predictions

31. 1. Descriptive
2. Predictive
3. Prescriptive
31
Phases of data analysis

32. 1. Show churn rate against time
2. Predict which customers will churn next
3. Suggest what to do about each customer
(e.g. propose to switch plan, send promotional offer, etc.)
32
Churn analysis

33. 1. Show returned goods against {type, customer segment}
2. Predict risk shopper will return goods
3. ?
33
E-commerce returns

34. “Suggest what to do about each customer”→ prioritised list of actions,
based on…
• Customer representation + context
• Churn prediction & action prediction
• Uncertainty in predictions
• Revenue brought by customer & Cost of actions
• Constraints on frequency of solicitations
34
Churn analysis

35. 35
Pricing optimisation
Again, from David Jones (@d_jones, see source)

36. Decide price given product and context…
• For several price candidates (within constrained range):
• Predict # sales given product, context, price
• Multiply by price to estimate revenue
36
Pricing optimisation

37. Decide price given product and context…
• For several price candidates (within constrained range):
• Predict 95%-confidence lower bound on # sales given
product, context, price
• Multiply by price to estimate revenue
37
Pricing optimisation

38. 1. Show past demand against calendar
2. Predict demand for [product] at [store] in next 2 days
3. Suggest how much to ship
• Trade-off: cost of storage vs risk of lost sales
• Constraints on order size, truck volume, capacity of people
putting stuff into shelves
38
Replenishment

39. • Context
• Predictions
• Uncertainty in predictions
• Constraints
• Costs / benefits
• Competing objectives ( trade-offs to make)
• Business rules
39
Decisions are based on…

40. 40
Who performs better?
+vs.
Star Wars: The Flat Awakens
by Filipe de Carvalho
vs.

41. 41
AI + Human perform better
+

42. 42
Human alone performs better: dexterity

43. 43
AI alone performs better: replenishment

44. Decisions are faster, cheaper, and better
44
AI alone performs better: replenishment
Again, from Lars Trieloff @trieloff (see source)
Decision Quality
Status Quo Predictive Prescriptive Automation
Decisionquality

45. 1. Descriptive analysis
2. Predictive analysis
3. Prescriptive analysis
4. Automated decisions
45
Beyond prescriptive analysis

46. • Spam filter → decide to skip inbox
• Autonomous Vehicles → decide who to kill
46
Autonomous decision-making systems
“Tool AI”vs“High-stakes autonomous AI”

47. 47
Autonomous Vehicles

48. • Morality in decision-making algorithm:
• Minimize loss of life
• Account for probabilities of survival, age of occupants…
→ optimal formula?
• Sacrifice owner?
• “People are in favor of cars that sacrifice the occupant to save other
lives—as long they don’t have to drive one themselves.”
48
Autonomous Vehicles

49. • Need wide acceptation to get adoption and provide benefit (e.g.
save lives with AVs)
• “The public is much more likely to go along with a scenario that aligns
with their own views”
• What will the public tolerate? → experimental ethics
• Similar issues whenever AI decides for us and impacts many
“Domain-specific/business rules”in decision making
49
High-stakes autonomous AIs

50. Role of APIs

51. 51
Communication between AIs
01000101101

52. Software components for automated decisions:
• Create training dataset from historical data (merge sources, aggregate…)
• Provide predictive model from given training set (i.e. learn)
• Provide prediction against model for given context
• Provide optimal decision from given contextual data, predictions,
uncertainties, constraints, objectives, costs
• Apply given decision
52
Separation of concerns

53. Software components for automated decisions:
• Create training dataset from historical data (merge sources, aggregate…)
• Provide predictive model from given training set (i.e. learn)
• Provide prediction against model for given context
• Provide optimal decision from given contextual data, predictions,
uncertainties, constraints, objectives, costs
• Apply given decision
53
Operations Research component

54. Software components for automated decisions:
• Create training dataset from historical data (merge sources, aggregate…)
• Provide predictive model from given training set (i.e. learn)
• Provide prediction against model for given context
• Provide optimal decision from given contextual data, predictions,
uncertainties, constraints, objectives, costs
• Apply given decision
54
Machine Learning components

55. Software components for automated decisions:
• Create training dataset from historical data (merge sources, aggregate…)
• Provide predictive model from given training set (i.e. learn)
• Provide prediction against model for given context
• Provide optimal decision from given contextual data, predictions,
uncertainties, constraints, objectives, costs
• Apply given decision
55
Predictive APIs

56. 56
Predictive APIs

57. The two phases of machine learning:
• TRAIN a model
• PREDICT with a model
57
Predictive APIs

58. The two methods of predictive APIs:
• TRAIN a model
• PREDICT with a model
58
Predictive APIs

59. The two methods of predictive APIs:
• model = create_model(‘training.csv’)
• predicted_output = create_prediction(model,
new_input)
59
Predictive APIs

60. Amazon ML
BigML
Google Prediction
PredicSis
…
60
Providers of REST http Predictive APIs

61. Going further

62. • Define desired and acceptable behaviour
→ objectives and constraints/bounds
• Monitor accuracy & bottomline
• Self-monitoring & anomaly detection
→ thresholds and fallbacks
62
Ensuring performance of autonomous AI systems

63. 63
Performance guarantees?
“construction worker in orange safety vest
is working on road”
95%-accurate scene description

64. 64
Performance guarantees
“black and white dog jumps over bar”
95%-accurate scene description

65. 65
Performance guarantees
“a young boy is holding a baseball bat”
95%-accurate scene description

66. 66
Performance guarantees
“a young boy is holding a baseball bat”
weapon
SIR, DROP THE WEAPON!

67. • Lars Trieloff:“Business reasons for automating decisions”
• Daniel Kahneman: “Thinking, Fast and Slow”
• Tom Dietterich:“Artificial Intelligence Progress”
• MIT Technology Review:“Why Self-Driving Cars Must Be
Programmed to Kill”
• Conference: PAPIs Connect
67
Learn more

70. • Free ML resources: louisdorard.com
• PAPIs updates: @papisdotio

71. @louisdorard
LIKED
IT?
THEN