Industry Disruptors: AI, Machine
Learning and Drones
Dr. Anand S. Rao – Innovation Lead, PwC Data& Analytics
Dr. Falko Kuester – Prof of Visualization & VR, UCSD
www.pwc.com/analytics
January, 2017
PwC CES-2017
Today’s discussion
2
1. The Future is Now - Trends & Drivers for AI, Machine Learning and Drones
2. From the art of the possible to pragmatic possibilities
3. Overcoming challenges and moving forward
PwC CES-2017
1. Future is Now
33
PwC CES-2017 4
Community Database
Cyber Infrastructure
Rapid Response
Mainstream Media
ANALYSIS
• Modeling & Simulation
• Machine Learning
• Data Fusion
• Visual Analytics
• Virtual & Augmented Reality
DISSEMINATION
• Field Testing
• Training
• Publishing
• 3D Printing
• Citizen Science
CURATION
• Data Storage
• Data Bases
• Meta Data Augmentation
• Blockchain
ACQUISITION
• Diagnostic Imaging & Sensing
• Analytical Diagnostics
• Communications
• Robotics / Drones
Data
PwC CES-2017 6
Creating an Ecosystem for Disruptive Technologies
PwC
Artificial Intelligence is a branch of computer science dealing with the
simulation of intelligent behavior in computers
Machine
Learning
Deep
Learning
Natural
Language
Processing
Deep Q&A
systems (or
Cognitive
Computing)
Natural
Language
Generation
Social Network
Analysis
Graph
Analysis
Robotics &
Drones
Sensors /
Internet of
Things
Knowledge
Representation
Simulation
Modelling
Visualization
Image
Analytics
Audio/Speech
Analytics
Machine
Translation
Virtual
Personal
Assistants
Recommender
Systems
Deep Causal
Reasoning
Topic Areas within Artificial Intelligence (non-exhaustive)
PwC CES-2017
2. From the art of the possible to pragmatic
possibilities
PwC CES-2017
Rating product “style” with image and text analysis
Project: Deep learning to identify vehicle features The Path to Value...
Automobile Images Make & Model Prediction Start with general Image
recognition model and stock photos
Use “transformations” to train for
auto
Combine with style ratings
to assess new designs…
Refine design based on preferences
by customer segment
Create generative model to assist
designers
9
PwC CES-2017
Simulating market adoption with virtual models and agents
Project: Simulate adoption of personal mobility solutions The Path to Value...
Simulate a million ‘consumer’
agents and their purchase choices
based on causal reasoning
Run over 200K go-to- market
scenarios to prescribe the right city,
pricing, and # of vehicles
Personal mobility as a service
disrupting the transportation
sector
Incorporate real-time sensor data
from city and vehicles
Vehicle Fleets
(Driverless, Electric,
Sharing)
Simulating demand, charging and utilization by
geography
Modeling demand for vehicle miles travelled
10
PwC CES-2017
Identification of materials and boundaries with drones
Project: Capture and assess drone images from construction environment The Path to Value...
OriginalCroppedImageModelOutput
Extend image analysis to object
detection and semantic
segmentation
Increase number of drone runs to
collect image time-series
Track volume of high value
materials usage
Reduce risks and enhance return
on capital investment projects
Add additional data types (e.g.,
thermal, etc.)
Key:
Background Trees
Asphalt
ConcreteCars
Reinforcement
11
PwC’s Data & Analytics
PwC has worked with a number of clients across all sectors to solve
challenging problems for them using AI
12
Example Recent PwC AI Projects
1
Aerial Image Analytics
for Construction
Applies deep
learning to images
of property (e.g.
cars) to automate
the assessment
of the damage
to assets
2
Smart Coolers
Uses a connected
network of sensors
to derive valuable
insights about their
surroundings and
activities of those
within it
3
NLP for Cancer
Research
Uses natural
language
processing and
question-answering
systems to improve
quality of patient
care
4
Personal Mobility
Uses agent-based
modeling to run
over 200k
strategies to create
a business unit
focused on ride
share and
driverless cars
5
Speech Recognition
Deep Learning
Applies a suite of
deep learning
algorithms to convert
call center data into
text, which can then
be analyzed using
NLP techniques
6
Financial Advisor
Support
Uses agent-based
modeling to
simulate household
financial security
under a range of
scenarios to
provide advice
7
Predictive
Maintenance for Oil
Wells
Uses semi-
supervised natural
language processing
to automatically
extract information
from commercial
leasing contracts
PwC’s Data & Analytics
Drone Video
13
https://pwc.mediaspace.kaltura.com/media/entryId/1_cra40b7z
PwC’s Data & Analytics Video: Parking Lot
Asset Tracking
PwC CES-2017
3. Overcoming Challenges & Moving Forward
15
PwC CES-2017
Five key success factors to derive maximum benefits
from AI, big data & analytics
Start from business
decisions
Demonstrate value through
pilots before scaling
Fail forward – test
and learn culture
Address ‘big data’ –
don’t forget ‘lean’ data
Blend intuition and data-
driven insights
PwC CES-2017
Machine Learning for Augmented Intelligence
https://pwc.mediaspace.kaltura.com/media/t/1_yee8khtl
For more information:
Anand S. Rao
(anand.s.rao@pwc.com)
Twitter: @AnandSRao
© 2016 PwC. All rights reserved. PwC refers to the US member firm or one of its subsidiaries or affiliates, and may sometimes refer to the PwC network. Each member firm is a separate legal
entity. Please see www.pwc.com/structure for further details..
Images from Google Deep Dream
For more information:
Falko Kuester
(fkuester@ucsd.edu)
chei.ucsd.edu

Industry Disruptors: AI, Machine Learning and Drones.

  • 1.
    Industry Disruptors: AI,Machine Learning and Drones Dr. Anand S. Rao – Innovation Lead, PwC Data& Analytics Dr. Falko Kuester – Prof of Visualization & VR, UCSD www.pwc.com/analytics January, 2017
  • 2.
    PwC CES-2017 Today’s discussion 2 1.The Future is Now - Trends & Drivers for AI, Machine Learning and Drones 2. From the art of the possible to pragmatic possibilities 3. Overcoming challenges and moving forward
  • 3.
  • 4.
  • 5.
    Community Database Cyber Infrastructure RapidResponse Mainstream Media ANALYSIS • Modeling & Simulation • Machine Learning • Data Fusion • Visual Analytics • Virtual & Augmented Reality DISSEMINATION • Field Testing • Training • Publishing • 3D Printing • Citizen Science CURATION • Data Storage • Data Bases • Meta Data Augmentation • Blockchain ACQUISITION • Diagnostic Imaging & Sensing • Analytical Diagnostics • Communications • Robotics / Drones Data
  • 6.
    PwC CES-2017 6 Creatingan Ecosystem for Disruptive Technologies
  • 7.
    PwC Artificial Intelligence isa branch of computer science dealing with the simulation of intelligent behavior in computers Machine Learning Deep Learning Natural Language Processing Deep Q&A systems (or Cognitive Computing) Natural Language Generation Social Network Analysis Graph Analysis Robotics & Drones Sensors / Internet of Things Knowledge Representation Simulation Modelling Visualization Image Analytics Audio/Speech Analytics Machine Translation Virtual Personal Assistants Recommender Systems Deep Causal Reasoning Topic Areas within Artificial Intelligence (non-exhaustive)
  • 8.
    PwC CES-2017 2. Fromthe art of the possible to pragmatic possibilities
  • 9.
    PwC CES-2017 Rating product“style” with image and text analysis Project: Deep learning to identify vehicle features The Path to Value... Automobile Images Make & Model Prediction Start with general Image recognition model and stock photos Use “transformations” to train for auto Combine with style ratings to assess new designs… Refine design based on preferences by customer segment Create generative model to assist designers 9
  • 10.
    PwC CES-2017 Simulating marketadoption with virtual models and agents Project: Simulate adoption of personal mobility solutions The Path to Value... Simulate a million ‘consumer’ agents and their purchase choices based on causal reasoning Run over 200K go-to- market scenarios to prescribe the right city, pricing, and # of vehicles Personal mobility as a service disrupting the transportation sector Incorporate real-time sensor data from city and vehicles Vehicle Fleets (Driverless, Electric, Sharing) Simulating demand, charging and utilization by geography Modeling demand for vehicle miles travelled 10
  • 11.
    PwC CES-2017 Identification ofmaterials and boundaries with drones Project: Capture and assess drone images from construction environment The Path to Value... OriginalCroppedImageModelOutput Extend image analysis to object detection and semantic segmentation Increase number of drone runs to collect image time-series Track volume of high value materials usage Reduce risks and enhance return on capital investment projects Add additional data types (e.g., thermal, etc.) Key: Background Trees Asphalt ConcreteCars Reinforcement 11
  • 12.
    PwC’s Data &Analytics PwC has worked with a number of clients across all sectors to solve challenging problems for them using AI 12 Example Recent PwC AI Projects 1 Aerial Image Analytics for Construction Applies deep learning to images of property (e.g. cars) to automate the assessment of the damage to assets 2 Smart Coolers Uses a connected network of sensors to derive valuable insights about their surroundings and activities of those within it 3 NLP for Cancer Research Uses natural language processing and question-answering systems to improve quality of patient care 4 Personal Mobility Uses agent-based modeling to run over 200k strategies to create a business unit focused on ride share and driverless cars 5 Speech Recognition Deep Learning Applies a suite of deep learning algorithms to convert call center data into text, which can then be analyzed using NLP techniques 6 Financial Advisor Support Uses agent-based modeling to simulate household financial security under a range of scenarios to provide advice 7 Predictive Maintenance for Oil Wells Uses semi- supervised natural language processing to automatically extract information from commercial leasing contracts
  • 13.
    PwC’s Data &Analytics Drone Video 13 https://pwc.mediaspace.kaltura.com/media/entryId/1_cra40b7z
  • 14.
    PwC’s Data &Analytics Video: Parking Lot Asset Tracking
  • 15.
    PwC CES-2017 3. OvercomingChallenges & Moving Forward 15
  • 16.
    PwC CES-2017 Five keysuccess factors to derive maximum benefits from AI, big data & analytics Start from business decisions Demonstrate value through pilots before scaling Fail forward – test and learn culture Address ‘big data’ – don’t forget ‘lean’ data Blend intuition and data- driven insights
  • 17.
    PwC CES-2017 Machine Learningfor Augmented Intelligence https://pwc.mediaspace.kaltura.com/media/t/1_yee8khtl
  • 18.
    For more information: AnandS. Rao (anand.s.rao@pwc.com) Twitter: @AnandSRao © 2016 PwC. All rights reserved. PwC refers to the US member firm or one of its subsidiaries or affiliates, and may sometimes refer to the PwC network. Each member firm is a separate legal entity. Please see www.pwc.com/structure for further details.. Images from Google Deep Dream For more information: Falko Kuester (fkuester@ucsd.edu) chei.ucsd.edu