SlideShare a Scribd company logo
Disclosure of this information requires written approval from Government Acquisitions, Inc.
G   AI for the Mission
PREM JADHWANI
CHIEF TECHNOLOGY OFFICER
PREM.JADHWANI@GOV‐ACQ.COM
PHONE: 703‐554‐3827
NOVEMBER 13, 2018
Disclosure of this information requires written approval from Government Acquisitions, Inc.
AGENDA
• AI – Current State and Future
• Choosing the Right AI Workloads
• AI Use Cases for Public Sector
• AI  Maturity  Model
• AI 7 Layer Technology Stack
• AI Roadmap to Success
• What is GAI doing to enable AI ?
Disclosure of this information requires written approval from Government Acquisitions, Inc.
3
AI / ML – Why Should We Care?
Disclosure of this information requires written approval from Government Acquisitions, Inc.
Artificial Narrow Intelligence (ANI)
• AI Stages: Execute specific focused tasks, without 
ability to self‐expand functionality
• Timing: Today
•Implications: Outperform humans in specific repetitive 
functions, such as driving, medical diagnosis, and 
financial advice
Artificial General Intelligence (AGI)
•AI Stages: Perform broad tasks, reason, and improve 
capabilities comparable to humans
• Timing: About 2040?
•Implications: Compete with humans, such as earning 
university degrees and convincing humans it is human
Artificial Super Intelligence (ASI)
•AI Stages: Demonstrate intelligence beyond human 
capabilities
• Timing: Soon after AGI
•Implications: Outperform humans, helping to achieve 
societal objectives or threatening human race
4
Artificial 
Intelligence
Future
Artificial 
Intelligence
Machine ability to 
understand, learn, and 
act on information and 
events, designed to 
augment, provide 
assistance to, or perform 
tasks independently 
from humans
Disclosure of this information requires written approval from Government Acquisitions, Inc.
Mundane Repetitive
Human Tasks & Routine
Processes (Workforce
Automation through RPA)
Large Diverse Data
Sets in a Data Lake
(Need to find Patterns
& Correlations using
Object Tracking)
Complex & Massive
Cyber Log Data
(Predictive Threat
Detection & Mitigation)
Large Scale IOT Data
from Autonomous
Sensors (Image
Classification for Smart
Surveillance & Biometrics)
AI helps Reduce Administrative
Burden, Resolve resource
allocation problems, Automate the
Workforce, Improve Worker
Productivity, Solve Complex
Mission Challenges
What Workloads are Suitable for AI ?
Disclosure of this information requires written approval from Government Acquisitions, Inc.
AI Use‐Cases for the Public Sector 
Disclosure of this information requires written approval from Government Acquisitions, Inc.
Key  Functions Replaced By Bots Using 
Robotic Process Automation (RPA)
7
Disclosure of this information requires written approval from Government Acquisitions, Inc.
AI Maturity Model –
Path to AI Optimized Agency
Descriptive 
Analytics
Diagnostic 
Analytics
Predictive 
Analytics
Prescriptive 
Analytics
Cognitive 
Analytics
8
Today
Emerging
Operational Analytics
Advanced Analytics
Insight
What Happened and Why
Hindsight
What Happened
Self‐Learning and Completing 
Automated Enterprise
Computerized Human Thought Simulation 
Actions Towards Autonomic Agency
Simulation‐Driven Analysis and 
Decision‐Making
Mature Data Lake
Foresight
What Will Happen, When, and Why Analytics Curve
AI
Disclosure of this information requires written approval from Government Acquisitions, Inc.
ARTIFICIAL INTELLIGENCE PROCESS FLOW
Feed Learner 
Various Data 
(e.g., structured and 
unstructured)
Align Appropriate 
Type of Learning 
System 
(e.g., supervised and 
unsupervised)
Present Results
(e.g., exploratory, 
predictive and 
classification)
Input Data
“Learning”
Machine Learning
System
Output Data
Disclosure of this information requires written approval from Government Acquisitions, Inc.
SECURITY 
PRIVACY 
GOVERNANCE 
ETHICS
FAIRNESS
Rapid Data Visualization & Decision Making 
Rich Data Visualization & Mission Enablement by Data Scientists, Data Analysts, AI Center of Excellence (COE).
Containerized GPU Accelerated AI Applications
GPU Databases, Satellite Imagery Fusion, Robotic Process Automation, Intelligent Image Recognition, Responsibility 
Determination Bots, User Behavior Analytics for Insider Threat Detection
Deep Learning User Software
Apache Arrow, Apache Spark, NVIDIA RAPIDS, NVIDIA DIGITS, Supervised and unsupervised Learning Software
Deep Learning Frameworks
TensorFlow, Caffe, Pytorch, MXnet, H2O, CNTK, SciKit, Theano, Apache SINGA, CUDA Cores
Data Ingestion & Preparation
High Speed Data Ingestion, Cleansing, Classification, Normalization, Tagging & Labelling, Indexing and 
Correlation, Data Modeling & Training
AI Enabled Smart Endpoints
Sensors, Drones, Smart Devices, Cameras, LIDAR, IOT Devices, Autonomous Endpoints 
AI Ready High Performance Infrastructure
Hybrid Cloud, High Performance Compute, Hyper Converged Infrastructure, Optimized Storage, GPU Accelerated HPC Purpose Built 
Appliances,, FPGA & ASICs, High Speed Communication Networks, Geo Distributed Data Lakes, Virtualization & Containers, SDN, 
Blockchain,  Object Storage, Next Gen Security Platforms.
7 Layers 
of AI 
Technology 
Stack
(Important Enablers 
Designed for 
Ultimate AI
Visualization)
7
6
5
4
3
2
1
Disclosure of this information requires written approval from Government Acquisitions, Inc.
Roadmap to Success for AI Enabled Agency
1. Identify 
The Problem
2. Get The 
Endorsement
3. Assess The 
Capabilities
4. Look For 
Partnership
5. Create The 
Right 
Mindset
11
How to succeed in your next AI project
Agencies that want to leverage AI in an effective way should consider the following:
1. Identify the problems that are 
preventing you from achieving 
your mission goals and focus your 
AI efforts on them in accordance 
with your specific industry trends 
and unique mission needs.
2. Define a clear ownership structure and 
obtain full support from the agency leadership 
& C‐level officers. Everyone should 
understand that starting the AI journey implies 
an adoption curve and that learning might be 
itself one of the greatest outcomes.
3. Acknowledge the 
internal capability gap to 
understand what your 
team is able to do and 
what not.
5. Spread a digital 
culture and prepare and 
educate all those in the 
agency who will be 
affected by the change.
4. Partner for capability and capacity: AI 
workforce is today in short supply and 
developing complex AI solutions in‐house 
can be highly costly and inefficient. In this 
context, the contribution of SMEs and AI 
Solution Providers can be invaluable.
Disclosure of this information requires written approval from Government Acquisitions, Inc.
Efficient 
workflow
Resources: 
AI tools & 
Scientists
Motivated 
leaders & 
AI Ready 
Workforce 
Clear 
business 
case & 
Use Case
Structured 
& Clean 
Data
Clear 
Outputs & 
Success 
Criteria
12
AI / ML – Keys to Effective Pilots
Successful AI Pilots
Disclosure of this information requires written approval from Government Acquisitions, Inc.
GPU ACCELERATED  DEEP LEARNING 
REAL TIME INDEXING & SEARCHING MACHINE DATA
FISMA
& RMF & 
FITARA 
Compliance
Cyber and 
Insider 
Threat 
Detection & 
Remediatio
n
IT 
Operational 
Analytics
Business 
Analytics
Application 
Analytics
IT & BUSINESS OPERATIONSCOMPLIANC & CYBER
Smart Edge 
Analytics
Federated 
Analytics
GPU 
Accelerated 
Smart 
Biometrics 
and Video 
Analytics
AI/ML/DEEP LEARNING 
SCALE–OUT  
INTEGRATED  
OPTIMIZED  
SOLUTION
MULTIPLE   
USE–CASES
GAI PROFESSIONAL SERVICES  +  PROJECT MANAGEMENT
WEB‐SCALE HYPERCONVERGED APPLIANCE HW
GAI HyperCAP 2.0 Reference Architecture
Disclosure of this information requires written approval from Government Acquisitions, Inc.
04
GWACs, NASA SEWP-V, ITES-
3H, CIO-SP3, DHS First
Source, OTAs
Contract Vehicles &
Procurement
02
AI Ready Technology Bundles,
Full Solution Stack, Solution
Design, Proof of Concepts
(POC)
AI Technology &
Solution Enablement
Professional Services, Project
Management, Technology
Lifecycle Management,
Subject Matter Experts,
Ongoing Support.
Professional Services
03
AI Workshop & Market
Research for Business
Decision Makers & Technical
Decision Makers
AI Educational Workshops
01
Artificial
Intelligence,
Machine Learning
& Deep Learning.
What is GAI Doing to Enable AI ?
Disclosure of this information requires written approval from Government Acquisitions, Inc.
Dedication is our DNA….“
”Mission is our mindset! 
Prem Jadhwani
Chief Technology Officer
Government Acquisitions Inc.
Prem.Jadhwani@gov‐acq.com
703‐554‐3827

More Related Content

Government Acquisitions Accelerated AI

  • 1. Disclosure of this information requires written approval from Government Acquisitions, Inc. G   AI for the Mission PREM JADHWANI CHIEF TECHNOLOGY OFFICER PREM.JADHWANI@GOV‐ACQ.COM PHONE: 703‐554‐3827 NOVEMBER 13, 2018 Disclosure of this information requires written approval from Government Acquisitions, Inc. AGENDA • AI – Current State and Future • Choosing the Right AI Workloads • AI Use Cases for Public Sector • AI  Maturity  Model • AI 7 Layer Technology Stack • AI Roadmap to Success • What is GAI doing to enable AI ?
  • 2. Disclosure of this information requires written approval from Government Acquisitions, Inc. 3 AI / ML – Why Should We Care? Disclosure of this information requires written approval from Government Acquisitions, Inc. Artificial Narrow Intelligence (ANI) • AI Stages: Execute specific focused tasks, without  ability to self‐expand functionality • Timing: Today •Implications: Outperform humans in specific repetitive  functions, such as driving, medical diagnosis, and  financial advice Artificial General Intelligence (AGI) •AI Stages: Perform broad tasks, reason, and improve  capabilities comparable to humans • Timing: About 2040? •Implications: Compete with humans, such as earning  university degrees and convincing humans it is human Artificial Super Intelligence (ASI) •AI Stages: Demonstrate intelligence beyond human  capabilities • Timing: Soon after AGI •Implications: Outperform humans, helping to achieve  societal objectives or threatening human race 4 Artificial  Intelligence Future Artificial  Intelligence Machine ability to  understand, learn, and  act on information and  events, designed to  augment, provide  assistance to, or perform  tasks independently  from humans
  • 3. Disclosure of this information requires written approval from Government Acquisitions, Inc. Mundane Repetitive Human Tasks & Routine Processes (Workforce Automation through RPA) Large Diverse Data Sets in a Data Lake (Need to find Patterns & Correlations using Object Tracking) Complex & Massive Cyber Log Data (Predictive Threat Detection & Mitigation) Large Scale IOT Data from Autonomous Sensors (Image Classification for Smart Surveillance & Biometrics) AI helps Reduce Administrative Burden, Resolve resource allocation problems, Automate the Workforce, Improve Worker Productivity, Solve Complex Mission Challenges What Workloads are Suitable for AI ? Disclosure of this information requires written approval from Government Acquisitions, Inc. AI Use‐Cases for the Public Sector 
  • 4. Disclosure of this information requires written approval from Government Acquisitions, Inc. Key  Functions Replaced By Bots Using  Robotic Process Automation (RPA) 7 Disclosure of this information requires written approval from Government Acquisitions, Inc. AI Maturity Model – Path to AI Optimized Agency Descriptive  Analytics Diagnostic  Analytics Predictive  Analytics Prescriptive  Analytics Cognitive  Analytics 8 Today Emerging Operational Analytics Advanced Analytics Insight What Happened and Why Hindsight What Happened Self‐Learning and Completing  Automated Enterprise Computerized Human Thought Simulation  Actions Towards Autonomic Agency Simulation‐Driven Analysis and  Decision‐Making Mature Data Lake Foresight What Will Happen, When, and Why Analytics Curve AI
  • 5. Disclosure of this information requires written approval from Government Acquisitions, Inc. ARTIFICIAL INTELLIGENCE PROCESS FLOW Feed Learner  Various Data  (e.g., structured and  unstructured) Align Appropriate  Type of Learning  System  (e.g., supervised and  unsupervised) Present Results (e.g., exploratory,  predictive and  classification) Input Data “Learning” Machine Learning System Output Data Disclosure of this information requires written approval from Government Acquisitions, Inc. SECURITY  PRIVACY  GOVERNANCE  ETHICS FAIRNESS Rapid Data Visualization & Decision Making  Rich Data Visualization & Mission Enablement by Data Scientists, Data Analysts, AI Center of Excellence (COE). Containerized GPU Accelerated AI Applications GPU Databases, Satellite Imagery Fusion, Robotic Process Automation, Intelligent Image Recognition, Responsibility  Determination Bots, User Behavior Analytics for Insider Threat Detection Deep Learning User Software Apache Arrow, Apache Spark, NVIDIA RAPIDS, NVIDIA DIGITS, Supervised and unsupervised Learning Software Deep Learning Frameworks TensorFlow, Caffe, Pytorch, MXnet, H2O, CNTK, SciKit, Theano, Apache SINGA, CUDA Cores Data Ingestion & Preparation High Speed Data Ingestion, Cleansing, Classification, Normalization, Tagging & Labelling, Indexing and  Correlation, Data Modeling & Training AI Enabled Smart Endpoints Sensors, Drones, Smart Devices, Cameras, LIDAR, IOT Devices, Autonomous Endpoints  AI Ready High Performance Infrastructure Hybrid Cloud, High Performance Compute, Hyper Converged Infrastructure, Optimized Storage, GPU Accelerated HPC Purpose Built  Appliances,, FPGA & ASICs, High Speed Communication Networks, Geo Distributed Data Lakes, Virtualization & Containers, SDN,  Blockchain,  Object Storage, Next Gen Security Platforms. 7 Layers  of AI  Technology  Stack (Important Enablers  Designed for  Ultimate AI Visualization) 7 6 5 4 3 2 1
  • 6. Disclosure of this information requires written approval from Government Acquisitions, Inc. Roadmap to Success for AI Enabled Agency 1. Identify  The Problem 2. Get The  Endorsement 3. Assess The  Capabilities 4. Look For  Partnership 5. Create The  Right  Mindset 11 How to succeed in your next AI project Agencies that want to leverage AI in an effective way should consider the following: 1. Identify the problems that are  preventing you from achieving  your mission goals and focus your  AI efforts on them in accordance  with your specific industry trends  and unique mission needs. 2. Define a clear ownership structure and  obtain full support from the agency leadership  & C‐level officers. Everyone should  understand that starting the AI journey implies  an adoption curve and that learning might be  itself one of the greatest outcomes. 3. Acknowledge the  internal capability gap to  understand what your  team is able to do and  what not. 5. Spread a digital  culture and prepare and  educate all those in the  agency who will be  affected by the change. 4. Partner for capability and capacity: AI  workforce is today in short supply and  developing complex AI solutions in‐house  can be highly costly and inefficient. In this  context, the contribution of SMEs and AI  Solution Providers can be invaluable. Disclosure of this information requires written approval from Government Acquisitions, Inc. Efficient  workflow Resources:  AI tools &  Scientists Motivated  leaders &  AI Ready  Workforce  Clear  business  case &  Use Case Structured  & Clean  Data Clear  Outputs &  Success  Criteria 12 AI / ML – Keys to Effective Pilots Successful AI Pilots
  • 7. Disclosure of this information requires written approval from Government Acquisitions, Inc. GPU ACCELERATED  DEEP LEARNING  REAL TIME INDEXING & SEARCHING MACHINE DATA FISMA & RMF &  FITARA  Compliance Cyber and  Insider  Threat  Detection &  Remediatio n IT  Operational  Analytics Business  Analytics Application  Analytics IT & BUSINESS OPERATIONSCOMPLIANC & CYBER Smart Edge  Analytics Federated  Analytics GPU  Accelerated  Smart  Biometrics  and Video  Analytics AI/ML/DEEP LEARNING  SCALE–OUT   INTEGRATED   OPTIMIZED   SOLUTION MULTIPLE    USE–CASES GAI PROFESSIONAL SERVICES  +  PROJECT MANAGEMENT WEB‐SCALE HYPERCONVERGED APPLIANCE HW GAI HyperCAP 2.0 Reference Architecture Disclosure of this information requires written approval from Government Acquisitions, Inc. 04 GWACs, NASA SEWP-V, ITES- 3H, CIO-SP3, DHS First Source, OTAs Contract Vehicles & Procurement 02 AI Ready Technology Bundles, Full Solution Stack, Solution Design, Proof of Concepts (POC) AI Technology & Solution Enablement Professional Services, Project Management, Technology Lifecycle Management, Subject Matter Experts, Ongoing Support. Professional Services 03 AI Workshop & Market Research for Business Decision Makers & Technical Decision Makers AI Educational Workshops 01 Artificial Intelligence, Machine Learning & Deep Learning. What is GAI Doing to Enable AI ?
  • 8. Disclosure of this information requires written approval from Government Acquisitions, Inc. Dedication is our DNA….“ ”Mission is our mindset!  Prem Jadhwani Chief Technology Officer Government Acquisitions Inc. Prem.Jadhwani@gov‐acq.com 703‐554‐3827