Exciting practical applications of scalable deep learning and image - - PowerPoint PPT Presentation

exciting practical applications of scalable deep learning
SMART_READER_LITE
LIVE PREVIEW

Exciting practical applications of scalable deep learning and image - - PowerPoint PPT Presentation

Exciting practical applications of scalable deep learning and image recognition in the cloud Georgi Kadrev @georgikadrev Imagga Technologies @imagga GTC16, April 5th 150+ companies 3,700+ developers all around the world 3Y 150Y 2B+ NEW


slide-1
SLIDE 1

Exciting practical applications

  • f scalable deep learning

and image recognition in the cloud

Georgi Kadrev @georgikadrev Imagga Technologies @imagga GTC’16, April 5th

slide-2
SLIDE 2

150+ companies 3,700+ developers all around the world

slide-3
SLIDE 3

150Y 3Y 2B+

NEW PHOTOS SHARED

EVERY DAY

(just to be forgotten in 30 minutes!)

slide-4
SLIDE 4

All the buzz words…

  • Artificial Intelligence
  • Machine Learning
  • Deep Learning
  • DNN
  • CNN
slide-5
SLIDE 5

Why not solving the problem using these via an auto-tagging API ?!

slide-6
SLIDE 6

As simple as:

  • 1. Submit an image
  • 2. Get list of tags or categories
  • 3. Do whatever you need with them
slide-7
SLIDE 7

The tags

  • physical objects: car, dog, computer, …
  • scenery and conceptual: office, job, family, …
  • related terms: car -> vehicle
slide-8
SLIDE 8

The categories

  • personal topics: events, beach, street view, …
  • safe/unsafe content
  • or customer defined:
slide-9
SLIDE 9

But who would need that… ?

slide-10
SLIDE 10

10

Unsplash: for search

  • Technologies used: auto-tagging
  • Impact: reduces/replaces manual tagging,

enhances search

slide-11
SLIDE 11

11

KIA: for user profiling and advertising

  • Technologies used: auto-tagging, color extraction
  • Impact: very precise personalised targeting
slide-12
SLIDE 12

12

Tavisca: for sorting out 25M hotel photos

  • Technologies used: custom auto-categorization
  • Impact: automates classification, improves

browsing experience

slide-13
SLIDE 13

Seoul National University: for waste sorting

Image source: Wikipedia

  • Technologies used: custom auto-categorization
  • Impact: automates pre-sorting of waste
slide-14
SLIDE 14

Eden: for organizing personal photos

http://edenphotos.io

slide-15
SLIDE 15

Are we close to human performance?

https://demo.algolia.com/clashOfTags/

51.1 % 32.5 %

slide-16
SLIDE 16

Thank You!

api@imagga.com twitter.com/imagga facebook.com/imagga

slide-17
SLIDE 17

Hipster Bar: to let only hipsters in his bar! :)

slide-18
SLIDE 18

How we’ve built it?

  • a lot of image data
  • precise model optimization
  • scalable infrastructure
slide-19
SLIDE 19

The data challenge and our feedback loop

slide-20
SLIDE 20

The specific content challenge and our customer-defined training

slide-21
SLIDE 21

The high throughput challenge and on demand infrastructure