DIGITAL

CAPABILITIES
Can you describe their Business Model?

Can you describe their Customer Experience?

Can you describe their Operating Model?

Can you describe the Technologies they are using?

Digital capability building blocks and transformation
Customer Digital New Business
Experience Operating Models
Models

Unified Data Platform

Digital Technologies

Digital Skills and Leadership

People and Culture

Target Operating model
Operating Models are the vehicles of delivering the value promised to the customer. So, this “How” part defines how you serve the
customer in the most efficient manner

The Partners & Alliances STRATEGY The Process capability shows how
capability defines which & GOALS various functions and locations
activities will be performed BUSINESS interact, how work is executed,
within the organization, by other MODEL
PARTNERS where hand-offs occur
parts of the parent groups,
& PROCESS
outsource partners, and by The Performance Metrics
ALLIANCES
external parties such as service capability shows the key
suppliers metrics and
measurement processes
The Assets Locations required to ensure that
& Resources ASSETS the Operating Model is
capability defines LOCATIONS KPIs effective and is delivering
which activities are RESOURCES DRIVE THE benefits
executed where, the OPERATING
assets used to MODEL DESIGN
execute the services, The Organization capability
outlines the organizational
and the resources
used including Data, structure, locations of where
and Information activities occur, the sourcing
of activities (external vs.
NETWORKS ORGANISATION
internal) and how changes to
The Networks and IT the model will be managed
capability outlines the
services, applications and
infrastructure supporting the PEOPLE The People Competency Skills capability
business process, and place IT COMPETENCY outlines the people implications in terms
requirements on data and SKILLS
of skills and behaviours required, the
information expected role distribution and the change
implications
Smile Recognition

1.Detect the first human face in the first image frame

2.and locate the twenty standard facial features position.
3.In every image frame, use optical flow to track the position of
left mouth corner and right mouth corner with accuracy of 0.01
pixels and update the standard facial feature position by face
tracking and detection.
4.If x direction distance between the tracked left mouth corner
and right mouth corner is larger than the standard distance plus
a threshold Tsmile, then we claim a smile detected.
5.Repeat from Step 2 to Step 3.

K Nearest Neighbors

Optical flow is the pattern of motion of objects which is usually used for
motion detection and object segmentation.
K Nearest Neighbors - CART
KNN is a non-parametric, lazy learning Supervised algorithm. Its purpose is to use a database in which the data
points are separated into several classes to predict the classification of a new sample point.
Lazy
it does not use the training data points to do
any generalization
Non-Parametric
it does not make any assumptions on the
underlying data distribution.
No Training Phase
Feature Similarity
How closely out-of-sample features
resemble our training set determines
how we classify a given data point:
1.A positive integer k is specified,
along with a new sample
2.We select the k entries in our
database which are closest to the
new sample
3.We find the most common
classification of these entries
4.This is the classification we give
to the new sample
•Credit ratings — collecting financial
characteristics vs. comparing people with
similar financial features to a database. By the
very nature of a credit rating, people who have
similar financial details would be given similar
credit ratings.
•Should the bank give a loan to an individual?
Would an individual default on his or her loan?
Is that person closer in characteristics to people
who defaulted or did not default on their loans?
•More advance examples could include
handwriting detection (like OCR), image
recognition and even video recognition.
How to understand Digital Technology?
Analytics
MODELS
• An Abstract representation of
IoT some process, be it a baseball
game, an election campaign,
cooking, hiring, firing, insurance
Block Chain etc

Has allowed us to • Simplifications

Cyber Security Govern via Algorithms
• Implicit and Explicit parts
AR/VR/MR
• Blind Spots

……….. • Identify Specific Relationships

………. • Inductively Explore

A MODEL NEEDS CONSTANT FEEDBACK AND UPDATION TO STAY RELEVANT

Peek into the past for the score!
Fair Isaac and Company ( Bill Fair and Earl Isaac) – introduced in 1989
FICO Score 300-850 (<650 is poor)

Experian, Equifax and TransUnion

+ Length of Job, Salary, Purchases,

Education, Zip Code, Voting
History, Health History,
= E-score!
Convictions

Unregulated Data

• Transparent, clear feedback loop

• Model updated if discrepancies are found
• Credit Industry is REGULATED
Can you model POLICING?
Inputs:
Part 1 Crimes: Homicide, Arson, Assault(+)
Part 2: Vagrancy, Panhandling, drug peddling(+++)

Assumption:
Low level crime creates an atmosphere of disorder
which breeds major crimes (Broken Windows Policing)

Anti Social behavior happens

mostly in impoverished
neighborhoods (targets poor)

These arrests then justify more

policing > same groups targeted
Self Exciting Point Process Models of Seismology is used > CRIMINALIZING POVERTY
Including Epidemic Type Aftershock Sequences
The Educational IMPACT!
College Ranking The Scale makes everyone shoot for the same targets as rankings
are relative
What goes into college ranking?
• SAT Scores Saudi Arabia’s King Abdulaziz Maths dept ( in 2 years) came behind
• Acceptance Rates Harvard in 7th place
• Student Teacher Ratios
• % graduation Students are investment portfolios with assets and liabilities
• % Freshman to Sophomore In order to optimize recruitment , specialized consulting firms have
• Living alumni who gave money sprung up
• ….. Noel-Levitz – ForecastPlus (segregates students), Rightstudent
• Zip Code (target the most promising candidates)
• Tuition and Fees
• Thomson Reuters Academic Citation In the wake of the feedback loop of the model, student application
• … consulting firms have sprung up > ThinkTank Learning (Steven Ma)
• …. predicts students chances and then sells package worth $$$ for
• 25% reputation score - survey different colleges > no admission money back

Cannot measure what student really learn > use proxies > easier to game
AI LEARNING MODELS
Knowledge

Inductive Deductive Feedback

knowledge based Explanation-Based

inductive Learning(EBL)
learning(KBIL) Relevance-Based Supervised UnSupervised
Learning(RBL)
Y = f(X) Clustering
Classification (Clustering Genes,
Regression Segmentation)

Semi-Supervised Reinforcement
Curation Rewards and
Photo archive (some are Punishment
labelled) AlphaGo Zero
Prior knowledge is essential
K-Means Clustering
K-means clustering is a type of unsupervised learning, which is used when you have unlabeled data (i.e., data without defined
categories or groups). The goal of this algorithm is to find groups in the data, with the number of groups represented by the variable K.

•Behavioral segmentation:
• Segment by purchase history
• Segment by activities on application, website,
Unsupervised or platform
• Define personas based on interests
• Create profiles based on activity monitoring
Unlabeled Data •Inventory categorization:
• Group inventory by sales activity
• Group inventory by manufacturing metrics
•Sorting sensor measurements:
Cluster into groups • Detect activity types in motion sensors
• Group images
• Separate audio
• Identify groups in health monitoring
Clusters •Detecting bots or anomalies:
• Separate valid activity groups from bots
• Group valid activity to clean up outlier
detection
Centroids
XiaoIce: THE NEXT ROMANTIC POET
Wings hold rocks and water tightly
In the loneliness
Stroll the empty
The land becomes soft…..

First AI-authored collection of poems published in China

CONSUMER IOT: AWS IOT BUTTON
WHAT IS IoT?
IoT VALUE CHAIN

Source: Arthur D. Little

IOT LAYERS AND PROTOCOLS
SI View
TCP/IP Model IoT Applications

Data Format Binary, JSON, CBOR

Design View
Application Layer CoAP, MQTT, XMPP, DDS
Front End - UTP
Transport Layer UDP
Enabling Infra
IPv6/IP Routing
Network/Internet Layer
Back End - ATP
6LoWPAN

IEEE 802.15.4 MAC

Link Layer
IEEE 802.15.4 PHY
Physical Radio
SOFTWARE DEFINED PRODUCT – VALUE MODELLING

Stochastic Model – encodes the value

Dry clothes as quickly as possible or for as
Time / Energy = f (fan speed, angular velocity of drum, heating element cheaply as possible while maintaining
temp, clothes temp, load weight, energy pricing) their integrity

Application - functionality
Interrogates model and actuates the hardware defined model

Analytics – interrogates the model

• Builds models
• Use descriptive analytics to report on cost savings
• Use predictive analytics to estimate when clothes will be dried and
when rollers will wear out
HOME AUTOMATION
IoT Design Methodology Steps – Home Automation example
Purpose and Requirements

Process Model Specification – Define Use Cases

Domain Model Specification – physical entities, Virtual entities, Devices, resources and Services

Information Model Specification – Define structure (relation, attributes) of all information in IoT system

Service Specifications – Map Process and Information Model to Services and define service specifications

IoT Level Specification – Define the IoT level for the system

Functional View Specification – Map IoT level to functional groups

Operational View Specifications – Define comms options, service hosting options, device options

Device and Component Integration – Integrate devices, develop and integrate the components

Application Development – Develop applications

Home Automation Example
Purpose and Requirements

Purpose: A Home Automation System that allows controlling of the lights in a home remotely using a web
application

Behavior: The Home automation system should have auto and manual modes. In auto mode, the system
measures the light level in the room and switches on the light when it is dark. In manual mode, the system
provides the option of manually and remotely switching on/off the light

System Management Requirement: The system should provide remote monitoring and control functions

Data Analysis Requirement: The system should provide local analysis of the data

Application deployment Requirement: The application should be deployed locally on the device, but should
be accessed remotely.
Home Automation Example
Process Model Specification – Define Use Cases

Mode?

Auto Manual

Light level Light State

Low High On Off

State: On State: Off State: On State: Off

Home Automation Example
Information Model Specification – Define structure (relation, attributes) of all information in IoT system

Virtual entity : Room

Virtual entity : Room Appliance
Entity Type: Room In room Entity Type: Appliance
ID: Room1 ID: Light1
Room#: Room1

Attribute: Light Level Attribute: Light State

Attribute Name: Attribute Name:
Light evel Lightstate
Attribute Type: level Attribute Type: state

Has light level Has light level Is In State

Is In State

Level: High Level: Low State: On State: Off

Home Automation Example
Domain Model Specification – physical entities, Virtual entities, Devices, resources and Services

User

Human
User Interacts with

Invokes/ Active Digital

Subscribes Artefact: App

Relates to Monitors
VE: Room PE: Room

Service
Associated with
Associated with VE: Appliance Relates to PE: Appliance Acts on

Exposes

Resource Associated with Hosts Device

Minicomputer

Attached to Attached to

NW OnDevice Sensor Actuator

Resource Resource LDR Relay
Home Automation Example
Service Specifications – Map Process and Information Model to Services and define service specifications

MODE Service:
Sets Mode to Auto or Manual
and retrieves Mode

STATE Service:
Sets State to on or off and
retrieves state
 Home Automation Example
Service Specifications – Controller Service Specification

Set Mode Input Schedule

Auto/Manual Mode: Interval: Every 5
Auto/Manual has input has schedule sec
State: On/Off
Set State
On/Off has end point EndPoint: /home/mode
Protocol: HTTP
Service EndPoint: /home/state
Name: Controller
Protocol: HTTP
Type: Native
has output

Name: Mode
Current Mode Output Type: REST
Auto/Manual State: On/Off
Name: State
Current State Type: REST
On/Off
Home Automation Example
IoT Level Specification – Define the IoT level (Level – 1) for the system

LOCAL
App CLOUD

REST Communication
REST Services

Database

Controller
Service

Resource

Device
LEVEL 1
Home Automation Example
Functional Groups

Application

Application Server Web App Database Server

Management Services Security

App Authentic
Native Services Web Services ation
Mgmt
Device Authoriza
Communication
Mgmt tion
Communication APIs Web APIs
DB Mgmt

Devices

Sensor Actuator Computing Devices

Home Automation Example
Mapping IoT Level to Functional Groups

6
6
6
4 5

4
1
3
2
1

2
Home Automation Example
Mapping Functional Groups to Operational View

Native Service: App Server: Django

Controller Service Web App: Django
Web Services: REST DB Server: my SQL
(M/S)

App Mgmt: Django

app mgmt Auth/Auth: Web App
DB Mgmt: Mongo DB DB
Device Mgmt:
Raspberry Pi
Comm API: REST API
Comm Protocol: Link
Computing Device: Layer: 802.1
Raspberry Pi Network Layer:
Sensor: LDR IPv4/IPv6
Actuator: Relay Switch Transport Layer: TCP
App Layer: HTTP
AMARA’S LAW

“We tend to overestimate

the effect of a technology in
the short run and
underestimate the effect in
the long run”
THANK YOU!