2. Agenda
Concepts of AI (Learning Algorithms) 30mins
Understanding AI (5 Spokes Framework) 60 mins
• Reasoning
• Communication
• Decision Making
• Interaction
Team Exercise 30 mins
6. The path to AI (the search for “Learnings”)
5
Learning
Focus on retaining correlations in intelligent
repositories and reapplication of inferences across
the company. Terms Like Deep Learning,
Representations and Autonomous Entity
3
Root Cause Analysis
Focus on correlation and inference building
with a specific context to the business. Terms
like Intelligent Reports and Smart Machines, .
1
Collection
The focus on Data acquisition Sensoring and
Governance to lay the foundation for Data.
Terms like Big Data and Data Lakes DashBoards
and Control reporting.
4
Simulation
Focus on feedback loops to forecasting results
and accuracy on predictions models. Terms life
Digital Twin, Lifeing Modelling.
What happened?
What is happening?
Why did it happened?
What will happen?
What
should
happen?
Descriptive Statistics
Inferential Statistics
2
Processing
Focus on getting Information out of data.
Terms like Data Analytics, Data Mining &
Analytic Reporting.
8. 8
Learn user’s behaviour based on voice commands
and can adjust settings automatically in
subsequent interactions
Target user with personalized products and
services ads based on their demographic profile,
search history, visited sites, liked social media
posts, etc.
Improve efficiency/quality in servicing customers by
using AI assistants (e.g., chatbots, robo-greeters in
bank branches and cardless ATM machines via facial
recognition)
Detect suspicious/fraudulent activities in network
and/or transactions using predictive analytics
Recommend music or videos based on user’s
historical consumption and preferences
Provide best driving routes, ETA, and/or match
drivers with riders based on historical and real-
time data
Smart Home Devices Media & Entertainment Navigation and Transportation
E-commerce & Targeted Ads Customer Service Assistants Security and Fraud Detection
What we see every day….
9. Machine Assisted Intelligence “Learning Algorithms”
data, software, hardware, and research
Larger and more
sophisticated datasets
Faster hardware
and better software
Larger and more
sophisticated models
Research breakthroughs for
training models
Increasing interconnectivity in the
research community (TensorFlow,
Theano, Caffe, Torch)
Increasing Availability, Use and Expertise
of Data Value, Governance and
Engineering
Increasing Maturity of
Cloud Computing and
Processing Power
Increased Development and
Experience of Learning Algorithms
(AlexNet, BatchNorm, DeepLearning)
https://spectrum-ieee-org.cdn.ampproject.org/c/s/spectrum.ieee.org/amp/stop-
calling-everything-ai-machinelearning-pioneer-says-2652904044
10. What are Learning Algorithms?
Artificial Intelligence is somewhat inaccurate as
systems are not intelligent alas, yet they learn.
It is a set of algorithms that learn from data to make
predictions .
11. What can Neural Networks (NN’s) or perceptron do?
11
MACHINE LEARNING
Dislike
Like
Provide training examples Distinguish likes from dislikes
Learn useful features
<Citrus-ness>
<Shape>
<
T
e
x
t
u
r
e
>
Define fields to describe fruit
<Sweetness= ? >
<Shape= ? >
<Colour= ? >
Pre-program rules
If <round> & <sweet>
Or if <red> & not <sour>
Or if <green> & <sour>
Provide input and get fixed output or error
<Apple> = Like
<Kiwi> = Like
<Banana> = ???
RULES-BASED SYSTEM
<Citrus-ness>
<Shape>
<
T
e
x
t
u
r
e
>
13. Let’s simplify the progression
Input
Input
Input
Hand-designed
rules
Output
Learned simple
features
Learned complex
features
Mapping from
features
Output
Hand-designed
features
Mapping from
features
Output
Rules-based systems
1960’s
Classic machine learning
2000’s
Deep learning
2020’s
Input
Random
Probability
Distribution
Output
Stochastic systems
1980’s
18. What are Learning Technologies Components?
Sensing the world
Perception
Learning from every
interaction
Communication
Optimizing to specific
outcomes
Decision making
Understanding
concepts & relations
Reasoning
Taking actions in the
world to achieve goals
Interaction
Computer Vision
Natural Language
Understanding & Generation
Forecasting and Operations
Research
Knowledge Graphs
and Representations
Reinforcement Learning
Answer questions about a scene
Determine if a growth is cancerous or not
Infer what happened to
characters in a story
Drive on city streets and highways
Identify objects in a scene
37. Interaction: Reinforcement Learning and its
Applications
Reinforcement Learning:
Dynamic Treatment Regimes
https://arxiv.org/pdf/1908.08796.pdf
Best approach to optimal decision making to Game
against a disease
https://opendatascience.com/deep-learning-research-review-week-2-reinforcement-learning/
40. Agenda
• Use case Workshop
• “Decision Needs Analysis”
• AI Solution Template
• Team Presentations
• Intro to Ethical AI
• Fairness. Ethics. Accountability. Transparency.
• Data vs Model
• AI Global Standards
45. Data Driven -Intelligence Theory
1
Learning
Focus on retaining correlations in
intelligent repositories and
reapplication of inferences across
the company. Terms Like Deep
Learning, Representations and
Autonomous Entity
3
Root Cause Analysis
Focus on correlation and inference
building with a specific context to
the business. Terms like Intelligent
Reports and Smart Machines, .
5
Collection
The focus on Data acquisition
Sensoring and Governance to lay
the foundation for Data. Terms like
Big Data and Data Lakes
DashBoards and Control reporting.
2
Simulation
Focus on feedback loops to
forecasting results and accuracy on
predictions models. Terms life
Digital Twin, Lifeing Modelling.
What happened?
What is happening?
Why did it happened?
What will happen?
What
should
happen?
Descriptive Statistics
Inferential Statistics
46. Set 1- Decision Needs Analysis
What critical business QUESTIONS do you need to
answer within your department?
Why is it important to know? What DECISION will you
make with this information?
Who will use
information?
How often will you use this
information?
1
What are the key predictors of loan defaults in 25 to
30 year Old Single Professionals?
This indicates the highest risk yet the highest reward
lending profile so we are looking to increase this sector of
our loans while running more predictions to better
underwrite the risk of defaults.
Loan Risk Assessor Weekly Loan Application Revie
What data is necessary to make this Decision Today? (Signals) What potential Models would we use?
(Prediction, Knowledge, Classification, Clustering, Forecast, Vision, NL)
1
What are the key predictors of loan defaults in 25 to 30 year Old
Single Professionals?
This indicates the highest risk yet the highest reward lending
profile so we are looking to increase this sector of our loans
while running more predictions to better underwrite the risk of
defaults.
47. Ste 2- Use case Template
se Case ID: F.21 Use Case Title: Loan Risk Predictor & Underwriting Intelligence
Data Maturity Level
How Mature are we to deliver use case?
Value Objective (Objective Key Result):
What is the estimated % Increase in Revenue, or Reduction of Risk % per loan category if you had the answer?
Driving Business Unit:
Contributing Business Unit: NA
Use Case Description
Describe what the use case would do and how would the insight be consumed?
Potential AI Technology / Learning Algorithm Techniques
Describe what kind of Learning Algorithm would you use for this use case and why?
Use Case Complexity
Gage Complexity
Feasibility Window
Gage Feasibility Window
Data Implications
What kind of Data would we need for this use case? Comment on Cleansing, Modelling, Annotation, etc.
Architectural / Infrastructure / Sensing Considerations / UX CX
Any new Sensing or existing architectures that will need to be considered?
Use Case Examples
Link use cases, research papers or offerings in the market that cover your idea.
Notes
Mention other considerations
49. Sensing the world
Perception
Learning from every
interaction
Communication
Optimizing to specific
outcomes
Decision making
Understanding
concepts & relations
Reasoning
Taking actions in the
world to achieve goals
Interaction
Computer Vision
Natural Language
Understanding & Generation
Forecasting and Operations
Research
Knowledge Graphs
and Representations
Reinforcement Learning
Answer questions about a scene
Determine if a growth is cancerous or not
Infer what happened to
characters in a story
Drive on city streets and highways
Identify objects in a scene
Mis-identification of Threat
Dis-advantaging Groups
Promoting Hate Speech
Incrementing Market Volatility
Pedestrian Fatality -Autonomous Vehicles
What are the implications to humans?
50. Managing Enterprise AI
Source: https://www.mckinsey.com/business-functions/mckinsey-analytics/our-insights/getting-to-know-and-manage-your-biggest-ai-risks
51. Fairness Ethics Accountability Transparency
FEAT principles were created
to guide better deployment of AI
Justifiability
Accuracy & Bias
Internal &
External Outcome
Explainability
Interpretability
Align to our Ethos
52. A I A S I A P A C I F I C I N S T I T U T E
AI & Fairness
Is the data used a fair representation
of reality.
Is our model having Unintended
Consequences, Systemic Issues?
54. AI & Transparency
Strong evidence on the accuracy of
the output for high-stake decisions.
Interpretation- why model output is counter-
intuitive & do I trust it?
55. Going beyond model debugging
What kind of action can we drive?
Actionable XAI
Explanations that drive actions
“I need to understand why a model
made a decision so I can complete a
regulatory audit report”
Human-AI Interaction for XAI
AI can interact and also learn from
human insight
“I want to easily see the reasoning so I
can correct errors and give better
feedback on what I want the output to
be”
End to End XAI
Moving away explaining a model to
explaining a business process.
“If I am assembling a car,
understanding every piece separately
is not sufficient. I must be able to audit
the assembly process.”
1 2 3
55
56. AI for Compliance
Right to an explanation if receiving
an adverse decision
Why did we decide an unpopular decision?
61. Data & Concept Governance
Source: https://twitter.com/AporiaAi/status/1406999597575254018/photo/1
62. IEEE P7003TM Standard for Algorithmic Bias
Considerations
•IEEE P7000: Model Process for Addressing Ethical Concerns During System Design
•IEEE P7001: Transparency of Autonomous Systems
•IEEE P7002: Data Privacy Process
•IEEE P7003: Algorithmic Bias Considerations
•IEEE P7004: Standard on Child and Student Data Governance
•IEEE P7005: Standard on Employer Data Governance
•IEEE P7006: Standard on Personal Data AI Agent Working Group
•IEEE P7007: Ontological Standard for Ethically Driven Robotics and Automation
Systems
•IEEE P7008: Standard for Ethically Driven Nudging for Robotic, Intelligent and
Autonomous Systems
•IEEE P7009: Standard for Fail-Safe Design of Autonomous and Semi-Autonomous
Systems
•IEEE P7010: Wellbeing Metrics Standard for Ethical Artificial Intelligence and
Autonomous Systems
Source: https://doi.org/10.1145/3194770.3194773
64. Competencies needed in an Enterprise for AI
64
5
Object Detection
& Monitoring
Robust detection, counting and tracking of
objects and people in a wide variety of
environments, enabling valuable
workflows in many real-world situations.
Image Classification
Video Alteration
Object Tracking and Counting
6 Optimization
AI-powered optimization boosts
the efficiency of business processes and
tasks, maximizing lift and ROI compared to
traditional optimization methods.
Routing
Disruption management
Re-optimization
7 Explainability
Making your AI explainable to users helps
drive adoption and lowers the barrier of
entry for users.
Technical Explanations
Bias Evaluation and Tracking
Sample-based Explanations
And more including...
● Recommender systems
● Assignment with constraints
● Association rule learning
● Human-AI interaction
● Image segmentation
● Image clustering
● Routing with constraints
● etc..
4 Time-Series Forecasting
Best-in-class AI-powered forecasting
provides better accuracy and can deliver
lift for a wide range of forecasting
scenarios.
Hybrid Forecasting Models
Statistical Forecasting
Deep-learning Forecasting Models
3 Anomaly Detection
Detect anomalies on objects in
natural environments and in various types
of data, allowing for near real-time
reaction.
Visual Anomaly Detection
Event-based Anomaly Detection
Anomalies in Forecasting Data
1 Text Extraction & Analysis
Accelerate the extraction of insights from
multiple forms of text, catching signals
that are easily overlooked by humans.
Text Summarization
Sentiment Analysis
Text Classification
2
Optical Character Recognition
Instantly transcribe text from natural
environments or digital documents,
reducing manual work and enabling
automation.
Documents
Handwritten Notes
Live Scenes and Video