Composite Tech and
Artificial Intelligence
Class 3
Mahesh Natarajan & Himanshu Warudkar
March 2024
Recap of Day 1 &2
• Digital Transformation is the next phase of business
transformation
• It goes beyond digital technologies
•Technology is critical for competitive advantage • It is no longer a question of “Should we”...it is about “How
•All companies are technology companies! Quickly”
•System of Record vs System of Engagement • Companies need to start with assessing their digital maturity
• Digitization and automation vs digitalization and digital
•Challenges with traditional IT transformation
•Move to customer-centric digital IT • Technology development methodologies that support rapid
transformation while maintaining customer centricity
•Need to integrate legacy with digital technologies(c) Himanshu Warudkar 2
while managing the two simultaneously
Havells
Who will be the largest Wipro
manufacturers of
residential electric Anchor
switches by 2030 ? Legrand
Others :
Industry 4.0 as a context
Mechanical Harnessed Mass Computers & Industry 4.0
Advantage Mechanisation Production Automation • Cyber-physical
• Simple Machines • Steam Power • Electricity • Electronic systems
• Draught Animals • Water-Power • Production Control Systems • Embedded
• Wind Power Innovation • Industrial systems
Automation • Internet of Things
09:15 PM 12:15 – 12:30 09:15 PM 4
Digital
Transformation Is
About Talent, Not
Technology*
You are the Bridge between Technology & Business!
* https://hbr.org/2020/05/digital-transformation-is-about-talent-not-technology
09:15 PM 5
Technology as
the language
for Digital
09:15 PM 6
Need for a
taxonomy
No consistent understanding in
practice
Confusion over what makes Digital
different from the past technologies
Common language to understand and
apply digital technologies
Data Compute
Analytic Infrastructu
s re
Data
Storage Digital
Security
Examine a
common
Data
landscape of an
Digital
Communicati
on end-to-end
Security Compute
Infrastructu
computer
Instrumentatio
re network
n / Embedded
Systems
User
Interfaces
8
Digital Technologies Taxonomy
Six Technology Categories / Elements
Digital Taxonomy Categories (Taxonomy Elements)
INSTRUMENTATI COMPUTATION COMMUNICATIO INTERFACES DATA DIGITAL
ON N MANAGEMENT SECURITY
Information Information Information Interaction Information Assurance
Gathering Processing Dissemination between Storage,
Machines/Human Retrieval,
s Preparation,
Analysis
First Order Capabilities
09:15 PM 9
Digital Technologies Taxonomy
Six Technology Categories / Elements
Digital Taxonomy Categories (Taxonomy Elements)
INSTRUMENTATI COMPUTATION COMMUNICATIO INTERFACES DATA DIGITAL
ON N MANAGEMENT SECURITY
• Sensors • Cloud • LAN/PAN • Web • Big Data • Information
• IoT Devices • Edge • WAN • Mobile • AI / ML Security
• Embedded • On-Premise • MANET • AR/VR • Analytics / • Privacy
Systems H/W, M/W, S/W • M2M Visualization • Ethical AI
Examples
09:15 PM 10
Digitization and
Second-Order
Capabilities
11
Second Order Capabilities result from
richer digital representation of the real
world
INFORMATION RICHNESS RESULTING CAPABILITIES FOR
DIMENSIONS BUSINESS
1. Information Scope Increase
2. Precision/Granularity of Information
1. Decision Making
3. Increased information shelf-life (retention) 2. Decision Execution
4. Real Time / Near-Real Time Information ◦ Automation
5. Greater Information Processing / Cognition ◦ Augmentation
6. Predictive Capabilities based on Information
12
Higher Order Capabilities - Overview
CONTEXT OF HIGHER ORDER LINE OF SIGHT TO BUSINESS
CAPABILITIES OUTCOMES
Decision Making & Decision Execution Contexts 1. Enhanced Cashflows
Business Processes: 2. Accelerated Cashflows
◦ CRM, SCM, Product Development, HRM, etc.
3. Reduced Risk / Volatility of Cashflows
Industry Domains
◦ Autonomous Vehicles, Bio-technology, Inventory
Management, Manufacturing, etc.
Intermediate / Process Objectives
◦ Turn-around-time, reduced attrition, asset utilization
Composite
Technologies
1. The Internet of Things (IoT)
2. Artificial Intelligence (AI)
3. Blockchain
14
IoT Definitions[1]
IoT Thing/Object:
• With regard to the Internet of things, this is an object of the physical world
(physical things) or the information world (virtual things), which is capable of
being identified and integrated into communication networks.
IoT Device:
• With regard to the Internet of things, this is a piece of equipment with the
mandatory capabilities of communication and the optional capabilities of
sensing, actuation, data capture, data storage and data processing.
Internet of things (IoT):
• A global infrastructure for the information society, enabling advanced services
by interconnecting (physical and virtual) things based on existing and evolving
interoperable information and communication technologies
https://www.youtube.com/watch?v=LlhmzVL5bm8
[1]
ITU, “Overview of the Internet of things,”
Landscape of “Things” within Iot
16
Anatomy of IoT
Process
• Sound • Sound
• Light • Light
• Computation
• Touch (Pressure, • Space (Location)
• Communication
Heat, Vibration, etc.) • Etc…
• Space (Location) • Data Handling
(Capture, Store,
Sense Process) Respond
(Sensors) (Actuators)
IoT Objects are typically special purpose. Their main function is NOT
to general purpose computation in contrast to typical computers
Commonalities & Differences?
Discussion
This Photo by Unknown Author is licensed under This Photo by Unknown Author is licensed under This Photo by Unknown Author is licensed under
CC BY-SA-NC CC BY CC BY-SA-NC
19
Blockchain
20
Block Chain Structure and Mechanism
• Blockchain is a decentralized, encrypted ledger technology with
inbuilt security mechanism which makes the data stored in a
blockchain practically tamperproof. The core aspects that
underpin the properties of a blockchain are,
Structure: Data stored in blocks with encrypted identification. [Hash, Data,
Hash of Parent Block]
Mechanism:
a. All changes to the data are preserved chronologically through cross
referencing blocks.
b. Multiple copies of the chains of data blocks stored in a peer-to-peer (P2P)
network of nodes.
c. Any change to the chain requires consensus among the nodes in the P2P
network.
https://www.youtube.com/watch?v=UWFJ1H702-E 21
Digital Building Blocks of Blockchain
22
AI
Research on Generative AI for Molecular Design & Synthesis:
https://www.youtube.com/watch?v=4CHuIyW1oNg
AI in biomedical research: https://www.uctv.tv/shows/A-Deep-Look-into-the-AI-Revolution-39246
Matterworks (company), video on their product providing AI-based quantitative metabolomics:
https://www.matterworks.ai/unlocking-scalable-quantitative-metabolomics-with-ai
MACHINE LEARNING -
A DEFINITION
Machine learning is the field of
study in Computer Science that
gives computers the ability to
learn without being explicitly
programmed.1
Attributed to Arthur Samuel
1
Park, C., Took, C. C., & Seong, J. K. (2018). Machine learning in biomedical engineering. 24
CONVENTIONAL PROGRAMMING VS
MACHINE LEARNING: A PERSPECTIVE
Example: Restaurant Billing Software
Mechanism Inputs:
Data:
Rules • Unit Price (p)
Conventional • Quantity of Purchase (q)
Software Answers • GST % (t%)
Program Rule:
Data • Price (P) = (1+t) x p x q
Output:
• Answer: Price Payable
25
CONVENTIONAL PROGRAMMING VS
MACHINE LEARNING: A PERSPECTIVE
Example: Object Detection Task
Mechanism Inputs:
Data:
Rules • Dog Pictures
A Machine • Other Pictures
Conventional
Learning
Software Answers Answer: For each picture
Software • Dog (or)
Program
Program • Not a Dog
Data Output:
Rules: A trained mathematical model to
identify dog pictures from picture inputs
26
MACHINE
LEARNING Supervised Learning
APPROACHES
Example: Object Detection Task
Inputs:
Unsupervised Learning
Data:
• Dog Pictures
• Other Pictures
Answer: For each picture
• Dog (or)
LABELS
• Not a Dog Reinforcement Learning Link
Link
27
Machine Learning Techniques
Supervised Learning is a machine Unsupervised Learning is a Reinforced Learning of algorithms
learning technique where data and machine learning technique where progresses through interaction with
correct labels are used to train a data without any labels are the environment, taking actions and
mathematical model which can be provided as input for the algorithm receiving feedback in the form of
used for prediction of labels based to discover patterns / classes rewards. The learning goal is to
on new sets of data. withing the data. maximize rewards.
Example: Predicting stock market Examples: Social Network Analysis, Examples: Recommendation engines
Market Segment Analysis
28
Natural Language
Processing
Speech Recognition
Natural Language Understanding
Natural Language Synthesis
Translation
Artificial General Intelligence (AGI / HLMI)
Link
• "Based on conjecture that every aspect of learning or any
other feature of intelligence can in principle be so
precisely described that a machine can be made to
simulate it.”
Artificial Superintelligence
Artificial • “Any intellect that greatly exceeds the cognitive
performance of humans in virtually all domains of
Intelligence interest. (Hypothetical)”
Continuum Current AI
• “Most current AI research, which focuses on specific
scientific and engineering problems and maintains a
distance to the cognitive sciences.”
30
Why study FluidAI?
One of India’s Top AI
companies, grounded in
technology, products
that work!
31
09:15 PM 12:15 – 12:30
Questions for Case in Session 3 (Fluid AI)
What business was Fluid AI in?
What is the core value proposition offered by Fluid AI to its customers vs its competition?
Why did Fluid AI focus on banking industry and that too on international banking clients to begin
with?
What other industries can Fluid AI expand into? Please state two other industries and elaborate
on some use-cases within these industries.
What business was Fluid AI in?
• Artificial Intelligence (AI)-based startup
• Product or Services company?
• AI platform company
What is the problem space?
• Challenges faced by organizations
• Competitive landscape
• Opportunities for scaling
• …?
Core value proposition
• Integrated business model Fluid AI had differentiated itself
from competitors by providing end-
• Explainable analytical / ML models and a Deep Learning
to-end AI solutions to companies
AI-based approaches
and end-users, right from front-end
• AI-based experiences and AI-based analytics user experience to making
predictions and providing back-end
• Build their own algorithms and a robust AI platform and
insights which would mean the
filing patents. ability to have a 360-degree view of
• Sector agnostic AI platform customer’s relationship with an
organization
• Outcome-based contracts
• Empowering users
• Cloud and In-premise model
What products is FluidAI into?
Conversational AI AI Analytics Workbench
How does one build a (great) product?
People Processes Technology
Key points to ponder upon
1. Product Launch Strategy
1. Creating the MVP
2. Product Roadmap
3. Key Technological Decisions
2. Marketing Strategy
1. Getting a foot in the door
2. Gaining credibility
3. Funding Strategy
1. Bootstrapping vs. Venture Funding
4. Beating the Competition
1. Core Value Proposition
2. Key Differentiators
5. Risk Appetite
1. Is failure an option?
Why international banking industry?
• Banking industry generates and consumes lots of data
• Many banks started differentiating themselves by leveraging data rather than simply
being custodians of money
• Regulatory requirements
• Competition from FinTechs
• Multiple use cases: front-office/mid-office/back-office
• International banks:
• Willingness to adopt new technologies
• Larger orders
• More profitable
What should FluidAI do next?
1. Move into new domains
2. Diversify product portfolio
3. Dilute equity to scale further
…?
Other Industries?
• Telecom
• Insurance
• Hospitality
• Healthcare
