DANIEL WILSON AND BEN CONKLIN Integrating AI with Foundation - - PowerPoint PPT Presentation

DANIEL WILSON AND BEN CONKLIN

Integrating AI with Foundation Intelligence

for Actionable Intelligence

INTEGRATING AI WITH FOUNDATION INTELLIGENCE

FOR ACTIONABLE INTELLIGENCE

“…in an arms race for artificial intelligence”

• Dr. Anthony Vinci, NGA

Cloud ML New Products

Azure ML Cognitive Services Azure Bot Service Watson ML Service DSX Cognitive Computing Amazon ML Cloud AI Cloud ML (TF) Leonardo SAP Analytics Cloud ML Server SPSS ML for z/OS

• Leonardo

SAP Predictive Analytics Office 365 PowerPoint Outlook .. Predictive Maintenance Targeted Marketing .. Demand Forecasting, Recommendations, Search, Merchandising Placement, Fraud.. Search, Ads, Gmail, Translation, YouTube, Maps.. Fraud Mgmt, SAPHIRE, S/4HANA, Fieldglass, Total Workforce Insight Cortana Assistant

• Alexa, Prime Air Delivery

Drones, Grocery Stores Google Assistant

• On-premise ML

Enhanced Products

MICROSOFT IBM AMAZON GOOGLE SAP

Neural Networks TensorFlow

CNTK Natural Language Processing Cognitive Computing GeoAI Computer Vision Dimensionality Reduction Object Detection Support Vector Machines

Object Tracking

Keras Theano scikit-learn T-SNE Random Forest

Machine Learning

Deep Learning

Artificial Intelligence

Caffe

Machine Learning

Deep Learning

Artificial Intelligence

1950’s 1960’s 1970’s 1980’s 1990’s 2000’s 2010’s

Machine Learning

Deep Learning

Artificial Intelligence

CNTK TensorFlow Theano

Natural Language Processing Video game behavioral AI

Robotics

Keras

Convolutional Neural Networks

IBM Watson scikit-learn

Computer Vision

Applications of Machine Learning

eCommerce Spam Filtering Fraud Detection Transportation Management Facial Recognition Cyber Intrusion Detection

Why Now?

• 3. Better Algorithms
• 1. More Data
• 2. More Compute

ML Value

Prediction

Automation Anomaly Detection Root-cause Identification

GIS users have been doing machine learning

GIS

Classification Clustering Prediction

Integration of Machine Learning and Deep Learning with GIS

GIS

Amazing Rate of Improvement

Image Recognition

IMAGENET

Pedestrian Detection

CALTECH

Object Detection

KITTI

Convolutional Neural Networks (CNNs)

Deep Learning

Real-Time Land Cover Classification

Microsoft Cognitive Toolkit

Use Artificial Intelligence to find system faults Predictive Maintenance

IBM Watson

Real-Time Object Recognition from Video

TensorFlow

GIS and Natural Language Processing Integration

“This is not either

human analysis or artificial intelligence,

it's got to be

some combination of the two.”

• Adm. Mike Rogers, Director, National Security Agency &

Commander of U.S. Cyber Command

Technology Drivers for Advanced Analytics

• 3. Better Algorithms
• 1. More Data
• 2. More Compute

Categories of Machine Learning in GIS

GIS

Classification Clustering Prediction

Machine Learning Tools in GIS

• Maximum Likelihood

Classification

• Random Trees
• Support Vector Machine

Clustering

• Empirical Bayesian

Kriging

• Areal Interpolation
• EBK Regression

Prediction

• Ordinary Least Squares

Regression and Exploratory Regression

• Geographically Weighted

Regression

• Spatially Constrained

Multivariate Clustering

• Multivariate Clustering
• Density-based Clustering
• Image Segmentation
• Hot Spot Analysis
• Cluster and Outlier Analysis
• Space Time Pattern Mining

Classification Prediction

Using ng the known wn to esti tima mate te the unknow nown Use Case: Accurately predict impacts of climate change on local temperature using global climate model data

Prediction

In ArcGIS: Empirical Bayesian Kriging, Areal Interpolation, EBK Regression Prediction, Ordinary Least Squares Regression and Exploratory Regression, Geographically Weighted Regression

The groupi uping ng of obser serva vatio ions ns based ed on simil ilarities ities of values es or locat catio ions ns Use Case: Given the nearly 50,000 reports of traffic between 5pm and 6pm in Los Angeles (from Traffic Alerts by Waze), where are traffic zones that can be used to elicit feedback from current drivers in the area?

Clustering

In ArcGIS: Spatially Constrained Multivariate Clustering, Multivariate Clustering, Density-based Clustering, Image Segmentation, Hot Spot Analysis, Cluster and Outlier Analysis, Space Time Pattern Mining

The process cess of decid cidin ing g to which h catego egory y an objec ect shoul uld d be assigne gned d based d on a traini ning ng dataset set Use Case: Classify impervious surfaces to help effectively prepare for storm and flood events based on the latest high-resolution imagery

Classification

In ArcGIS: Maximum Likelihood Classification, Random Trees, Support Vector Machine

Integration of Machine Learning and Deep Learning with GIS

GIS

Demo: Transfer Learning

Enterprise Approach to Machine Learning

Leveraging Geography for Improved Understanding

Distributed Analytics

Vector Raster Real-Time

Analyst Tools

Modeling Data Conditioning and Management Visualization and Exploration

Vector Unstructured Text Imagery Data Lakes Foundation Data Point Clouds Sensor Feeds

Data Conditioning

Extract, Transform, Enrich, Georeference, Validate Making Data Ready for Analysis And Use in Apps

Visualization and Exploration

Exploratory Data Analysis Visualization

Modeling and Spatial Analytics

Develop and Capture new tradecraft

Python

Data Science Spatio-Temporal Modeling

Analytic Services

Machine Learning & Artificial Intelligence

Real-Time Analytics Workflow

Real-Time Analytics Situational Awareness Analysis Alerting Big Data Archive Visualization Collection Contextualization

• GeoFencing
• Aggregation
• Detection
• Filter

Raster Analytics Workflow

Dynamic Image Processing Collection Visualization Change Detection Feature Extraction

• Ortho-on-the-fly
• Classification
• Feature Extraction
• Mosaicking

Vector Analytics Workflow

Classification, Clustering, Prediction

Big Data Archive Pattern-

• f-Life

Link Analysis Trend Analysis

• Space-Time
• Hot Spots
• Density
• Proximity

Big Data Analytics Predictive Analytics

Demo: Event Prediction

Applying to Intelligence Problems

Research Search Monitor Discover

Known Unknown Known Unknown

Locations and Targets Behaviors and Signatures

The Intelligence Cycle

Requirements Tasking Collection Processing Exploitation Dissemination

Modern Emphasis Traditional Emphasis

Leveraging Foundation Intelligence

Cultural Data Landscape Data Social Data Observations

Observe Orient Decide Act

Actionable Intelligence

OODA loop Speed is the key Speed is relative

Certainty

0% 100%

Time & Resources

Data Intelligence

X

Certainty

0% 100%

Time & Resources

Data

X

Intelligence

Reduce Time to Action

Impacts on Entire Organization

Implementing Analytic Platform

Plan for Evolving Structure Collaborate with Industry Prepare Infrastructure Enhance Analyst Tradecraft Focus on Verification and Validation