Axa Assurance Maroc - Insurer Innovation Award 2024
Deep Learning In Industries
1. 10/20/2017 Women in Big Data Event Hashtags: #IamAI, #WiBD
Oct 18th AI Connect Speakers
WiBD Introduction & DL Use Cases
Renee Yao
Product Marketing Manager,
Deep Learning and Analytics
NVIDIA
Deep Learning Workflows (w/ a demo)
Kari Briski
Director of Deep Learning
Software Product
NVIDIA
Deep Learning in Enterprise
Nazanin Zaker
Data Scientist
SAP Innovation Center Network
3. Agenda
AI Connect
• 6:00-7:00pm – Registration and Networking
• 7:00-7:15pm – “WiBD Introduction & DL Use
Cases”, Renee Yao, Product Marketing
Manager, Deep Learning and Analytics, NVIDIA
• 7:15-7:45pm – “Deep Learning Workflows
(with a live demo)”, Kari Briski, Director of
Deep Learning Software Product, NVIDIA
• 7:45-8:15pm – “Deep Learning in Enterprise”
by Nazanin Zaker, Data Scientist, SAP
Innovation Center Network
• 8:15-8:30pm - Wrap-up & Giveaways
February Apache Hadoop
Training @ Cloudera
May Apache Drill and
Apache Spark @ MapR
June Career Empowerment
@ Andreessen Horowitz
June @ Spark Summit
June @ Hadoop SummitMarch @ Strata+Hadoop
World SJ
10/20/2017 Women in Big Data Event Hashtags: #IamAI, #WiBD
4. 10/20/2017 Women in Big Data Forum
Be Part of The Solution
Become a member or a sponsor
• Website: womeninbigdata.org
• LinkedIn: “Women in Big Data Forum”
• Meetup: meetup.com/Women-in-Big-Data-Meetup/
• Twitter: @DataWomen
• Video: https://www.youtube.com/channel/UCOaMT7A9SVkeBdvYNxiITVA
Join us
Event Hashtags: #IamAI, #WiBD
5.
6. Renee Yao | NVIDIA Product Marketing Manager, DL and Analytics
Oct 2017 | Women in Big Data Meetup, AI Connect
AI CONNECT
8. 8
GPU COMPUTING AT THE HEART OF AI
New Advancements Leapfrog Moore’s Law
Performance Beyond Moore’s Law Big Bang of Modern AI
Twitter: @ReneeYao1
9. 9
DEEP LEARNING INDUSTRIES
Twitter: @ReneeYao1
Fashion
RETAILHEALTHCARE MANUFACTUREAGRICULTURE
$
FINANCIAL SERVICES RECYCLYINGINSURANCE
AI Now. For Every Industry.
10. 10
Twitter: @ReneeYao1
Best Service Offering Digital Assistance Account Security
BREAKTHROUGH CUSTOMER SERVICES
“ Future advancements in machine learning will unlock
opportunities for us to create breakthrough consumer experiences
in ways we can’t even imagine today..”
— Adam Wenchel, VP of AI, Capital One
BREAKTHROUGH CUSTOMER SERVICES
11. 11
Twitter: @ReneeYao1
AI ENABLES AUTOMATED
PROPERTY UNDERWRITING
• Geospatial Imagery, Computer Vision and Machine Learning
• A Living Database of The Most Accurate and Up-to-date Property Data
• Features: Roof Condition, Roof Geometry, Roof Covering Material, Solar Panels, Pool Enclosure, etc
• CUDA + CUDNN + NVIDIA GPUs
https://capeanalytics.com/
Twitter: @ReneeYao1
12. 12
Twitter: @ReneeYao1
DEEP LEARNING FOR
PORTABLE
DIAGNOSTICS
Portable Diagnostic Device that
Enables Rapid Analytics from a Drop of
Blood
Problem:
$100 billion wasted on treating
diseases diagnosed late
From a drop:
• WBC Counts
• Infection
• Bacterial vs. Viral
• Leukemia
• Immuno-suppression
• Heart Attack
Twitter: @ReneeYao1
13. 13
Twitter: @ReneeYao1
DEEP LEARNING FOR
PORTABLE
DIAGNOSTICS
Frameworks:
Keras and TensorFlow
Learning:
Supervised – base classifier
Unsupervised - labeling
Reinforcement – labeling small batch of
data
Outcome:
• In clinical validations, the Athelas
device has achieved 100% Clinical
Range accuracy for White Blood Cell
Counts.
• On track to 10,000 devices deployed
by end of the year
Twitter: @ReneeYao1
14. 14
Twitter: @ReneeYao1
$25 Billion Spent Each Year 3 Billion Pounds of Herbicides 250 Species of Resistant Weeds
BREAKTHROUGH CUSTOMER SERVICES
With the rise of herbicide-tolerant weeds, there are fewer and fewer effective solutions.
Farmers around the world need a new way to address the weed control challenge.
DEEP LEARNING FOR AGRICULTURE
15. 15
Twitter: @ReneeYao1Twitter: @ReneeYao1
Intelligent Weeding in Cotton
Computer Vision & ML
for Seed & Spray Equipment
• 10x unlocked chemical options to fight resistant
weeds when not spraying crops
• 90% lower chemical costs when selectively
applying herbicide to weeds only
• $2.5B potential reduction in global herbicide use
with sustainable weed control
DEEP LEARNING FOR AGRICULTURE
Make Every Plant Count
Verify and LearnSense and Decide Act
6-8 mph 8-12 rows 0-12 inches
speed of tractor,
operated by 1
person, as it pulls
the machine
through the field
of cotton
covered
simultaneously
in one pass by a
30-foot machine
range of cotton
sizes for which
See & Spray can
be used
16. 16
Twitter: @ReneeYao1Twitter: @ReneeYao1
DL FOR RECYCLING
AI | Computer Vision | Robotic Revolution
• 1.9 billion tons of waste are generated each
year
• Only 13% is recycled or recovered
• While Wall-B was able to perform 20 picks per
minute and could only recognize PET plastic
bottles, Max-AI can handle a much wider range
of items and at up to four times the previous
rate. Previous robotic solution underperformed
human capabilities, while Max-AI exceeds them.
• Save 80 tons of CO2 or 625 petrol barrows
yearly
17. 17
Twitter: @ReneeYao1Twitter: @ReneeYao1
DL FOR MANUFACTURE
AI | Computer Vision | Robotic Revolution
Upload Parts
Add to Estimation
Match Parts w/ Suppliers
Track Your Product
Receive Parts
“We were drawn to the idea of
on-demand manufacturing and
on-demand fabrication as ways
to reduce overhead and
simplify operations … With
MakeTime, it’s easy. You just
upload your parts, set your
specifications and you’re good
to go.”
Nazareth Ekmekjian,
Piaggio Fast Forward
18. 18
Twitter: @ReneeYao1Twitter: @ReneeYao1
DEEP LEARNING FOR FASHION SPOTTING
Deep Learning Powered Trends Spotting and Analysis for Fashion and Beauty
-Financial Times, July 2017
DL for Better Generalization Details
Model Architectures:
- In-house Architecture Based on
ResNet and DenseNet
- In-house Training Methods +
Transfer Learning
Frameworks:
• Keras: Fast Prototyping and
Classification Tasks
• TensorFlow: Flexibility for
More Complex Models
(Detection, Segmentation,
Metric Learning)
19. 19
Twitter: @ReneeYao1Twitter: @ReneeYao1
DEEP LEARNING FOR FASHION SPOTTING
Deep Learning Powered Trends Spotting and Analysis for Fashion and Beauty
-Financial Times, July 2017
Image Pipeline Details
Training Data:
• Classification/Localization/Se
gmentation: ~1000 training
examples per class
• 40k in-house images
Training: 8 TitanX Pascal
Inference: AWS GPU Instance
Results:
• 96% Accuracy in Precise
Localization and
Segmentation of Handbags
and Clothes
20. 20
Twitter: @ReneeYao1
SELECTED DEEP LEARNING USE CASES
Twitter: @ReneeYao1
Fashion
RETAILHEALTHCARE MANUFACTUREAGRICULTURE
$
FINANCIAL SERVICES RECYCLYINGINSURANCE
AI Now. For Every Industry.
21. 21
Twitter: @ReneeYao1
AI RESOURCES
Deep Learning Institute
Training Material
Hands-on Labs
Self-paced Courses
Nanodegree Programs
nvidia.com/DLI
Two Days To A Demo
Getting Started with AI Is
Easier Than You Think.
Develop Your First Demo
with Jetson
developer.nvidia.com/
embedded/twodaystoademo
Inception
Access to NVIDIA Tech
GPU and AI Experts
Global Marketing/Sales
GPU Venture Introduction
nvidia.com/inception
GTC
GTC Is the Largest and
Most Important Event Of
The Year For GPU
Developers
gputechconf.com
AI RESOURCES
Editor's Notes
There are no simple answers…
…but there is much companies can do to advance gender-equity in Computing and Big Data.:
Emphasize that diversity is a valid—and valuable—goal for business.
Introduce diversity initiatives and policies.
Expand recruitment.
Encourage women to speak out (lean in) and pursue leadership roles.
Recognize what motivates women (and incentivize it).
Foster mentors and role models.
And much, much more.
Be part of the solution: Join us today at www.womeninbigdata.org.
Home insurance companies must provide quotes to prospective customers quickly and accurately, but available data is either outdated or can take days to obtain. Cape Analytics uses GPUs to train deep learning models on frequently updated geospatial imagery to detect property features at a massive scale, creating a more accurate and up-to-date data source. Now, insurers can automatically generate accurate quotes for customers—enabling a better understanding of their risk profile, improving efficiency, and providing a better customer experience.