Toward Collaboration: Government Analytic Architecture Survey

November 2016Dr. Ransom Winder, MITREMr. Joseph Jubinski, MITRE

Dr. Angela O’Hanlon, MITRE

The views, opinions and/or findings contained in this report are those of The MITRE Corporation and should not be construed as an official government position, policy, or decision, unless designated by other documentation.

© 2016, The MITRE Corporation. All rights reserved.

Approved for Public Release; Distribution Unlimited. Case Number 16-4256

Introduction

• Established practice vs emerging capability?

• Differences?

• What drives preferences?

• What are the foundational elements of a reference implementation for collaboration?

Analytics and Collaboration

• Analytics: software that discovers patterns in data

• Ideally: data, analytics, and artifacts all equally accessible

• Typically: data & analytics walled off • Proprietary intellectual property

• Sharing artifacts is more likely

Data Analytic Results

AnalyticAnalytic

ResultsArtifactsData

Data

Ideally Typically

Artifacts

Artifacts

Artifacts

Analytic

Analytic

Data

Data

Data Analytic

Government: Defense/Intel and Civilian Agencies

Analytic Architectures in Government:Defense and Intelligence

?

• Established enterprise-wide architectures

• Efforts to refresh existing architectures

• Smaller projects for limited communities

Analytic Architectures in Government:Defense and Intelligence, Implementation Issues

Cloud

MobileDevice

Cloud

MobileDevice

Ideal Real World Conditions (e.g. Battlefield)

Data

Data

DataUseCase

Use Case

UseCase

IP

Intel Consumers

Defense and Intelligence Enterprises

DoD JIE – Department of Defense Joint Information Environment

IC ITE – Intelligence Community Information Technology Enterprise

DI2E – Defense Intelligence Information EnterpriseDiagrams derived from:http://www.afei.org/PE/4A07/Documents/DI2EPlugfest2014-JimMartinBrief-v0.pdf

DCGS

COCOM JIOC

IC ITE JIE

DI2E

Distributed Common Ground System

DCGS –Distributed Common Ground System

Distributed Common Ground System

AF DCGS – Air Force Distributed Common Ground System

Distributed Common Ground System

TCRI – Tactical Cloud Reference Implementation

Distributed Common Ground System

DIB – DCGS Integration Backbone

Analytic Architectures in Government:Civilian Agencies

• Important points• Analytic architectures can fit these contexts, but…

• Architectures must conform to legal restrictions / protections

• Restrictions dictate how data may be used

Centers for Medicare & Medicaid Services (CMS)

• Some enterprise-wide capabilities for data• Integrated Data Repository

• Chronic Condition Warehouse

• Big data sets• 4.5 million claims daily [1]

• Big budget • FY2015: $984.5 billion [2]

[1] https://www.cms.gov/Research-Statistics-Data-and-Systems/Computer-Data-and-Systems/HIGLAS/index.html[2] https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/CMS-Fast-Facts/index.html

Centers for Medicare & Medicaid Services (CMS)

Requirements

• Health Insurance Portability and Accountability Act (HIPAA)

• Affordable Care Act (ACA)

• Analytic expediency and integrity

• Decisions must be understandable and justifiable

However, predictive analytics have aided in catching fraud [1]

[1] https://blog.cms.gov/2016/07/20/42-billion-saved-in-medicare-and-medicaid-primarily-through-prevention/

Department of Veterans Affairs

• VA promotes an open source community • Open Source Electronic Health

Record Alliance (OSEHRA) [2]

• Gives third parties a baseline for interacting

• Activity: Build an ontology to influence future medical record schemas [1]

• Goal: Demonstrate the value of reasoning for health care

[1] Stoutenberg et al, "Developing a Clinical Care Ontology for the Veterans Health Administration“[2] www.osehra.org

Towards an Analytic Architecture for Industry and Government

Some orgs need an enterprise-wide analytic architecture

Others need individual analytic architectures for specific use cases

…But there is no established ubiquitous solution Analysis Exchange Model

1. Hub for collaborative analytic activity

2. Allows fusion across different sources

3. Anticipates contributions from across Industry

SourceMachine

OBP

ABI

Object Based Production &Activity Based Intelligence

• Object Based Production (OBP)• Aggregate knowledge into objects

• Activity Based Intelligence (ABI)• Reveal the underlying patterns

OBP

ABI

OutputHuman

As expressed in:https://info.publicintelligence.net/DIA-ActivityBasedIntelligence.pdf (Slide 8)

Source

Source

Source

Source

AnalystObject

Source

Source

Source

Source

Analyst

Object Based Production Flow for Intelligence Discovery

Previous approach OBP approach

As expressed in:https://info.publicintelligence.net/DIA-ActivityBasedIntelligence.pdf (Slide 4)

Analysis Exchange

Model

Source

Source

Source

Source

Object

Source

Source

Source

Source

Toward an Analytic Architecture for Industry and Government

OBP approach Our approach

Analytic Artifacts

Analytic Artifacts

Analytic Artifacts

Analytic Artifacts

Key Elements of an Analytic Architecture

Orchestrating exchange activityA “lingua franca” for exchangeAdaptation and fusion services

Conclusions

A push exists toward a comprehensive analytic architecture across and beyond their enterprise

Protective legal barriers must be considered

• Analytic architectures across various government agencies• Different needs and objectives

• Different scales of data

• Common need to derive valuable knowledge or intelligence

• No settled solutions for any agency

• Right time to consider alternatives• Foster collaboration between

government and industry • Mutual benefit• Collaboration can provide new

capabilities

Conclusions

• Roundtable’s proposal: Analysis Exchange Model• Hub for collaboration• Flexible for different use cases

Thank you

Questions?