SlideShare uma empresa Scribd logo
1 de 20
Proprietary and ConfidentialProprietary and Confidential
Proprietary and ConfidentialProprietary and Confidential
Artificial Intelligence for Mobile Visual Search
Catherine Ulrich, CPO
Proprietary and ConfidentialProprietary and Confidential
Proprietary and ConfidentialProprietary and Confidential
Proprietary and ConfidentialProprietary and Confidential
Building a Neural Network
State of the art computer vision techniques were researched, prototyped, then
iterated on until the service began to take form.
As the team iterated and iterated, the similarity results improved.
Let’s take a journey...
Source: 2012, Krachevsky et al
Proprietary and ConfidentialProprietary and Confidential
Training the model
Proprietary and ConfidentialProprietary and Confidential
Proprietary and ConfidentialProprietary and Confidential
Proprietary and ConfidentialProprietary and Confidential
Proprietary and ConfidentialProprietary and Confidential
Proprietary and ConfidentialProprietary and Confidential
Proprietary and ConfidentialProprietary and Confidential
Proprietary and ConfidentialProprietary and Confidential
Proprietary and ConfidentialProprietary and Confidential
Proprietary and ConfidentialProprietary and Confidential
Text
Proprietary and ConfidentialProprietary and Confidential
Proprietary and ConfidentialProprietary and Confidential
Proprietary and ConfidentialProprietary and Confidential
How is it performing
Proprietary and Confidential
uThank y
Proprietary and Confidential
uThank y

Mais conteúdo relacionado

Destaque

Destaque (10)

Joachim von Haenisch, CEO of KYC-Exchange
Joachim von Haenisch, CEO of KYC-ExchangeJoachim von Haenisch, CEO of KYC-Exchange
Joachim von Haenisch, CEO of KYC-Exchange
 
Presentation from BKM, Turkey
Presentation from BKM, TurkeyPresentation from BKM, Turkey
Presentation from BKM, Turkey
 
Why distributed ledgers
Why distributed ledgersWhy distributed ledgers
Why distributed ledgers
 
Fintech Innovation - Sw7 Innotribe Webber Wentzel
Fintech Innovation - Sw7 Innotribe Webber WentzelFintech Innovation - Sw7 Innotribe Webber Wentzel
Fintech Innovation - Sw7 Innotribe Webber Wentzel
 
Mobile, Fintech, and Banking Phase II
Mobile, Fintech, and Banking Phase IIMobile, Fintech, and Banking Phase II
Mobile, Fintech, and Banking Phase II
 
How to Envision the Digital Technology Strategy Industry Landscape
How to Envision the Digital Technology Strategy Industry LandscapeHow to Envision the Digital Technology Strategy Industry Landscape
How to Envision the Digital Technology Strategy Industry Landscape
 
The Omnichannel Opportunity
The Omnichannel OpportunityThe Omnichannel Opportunity
The Omnichannel Opportunity
 
London leads growth in RegTech investments. Research by FinTech Global, Jan 2017
London leads growth in RegTech investments. Research by FinTech Global, Jan 2017London leads growth in RegTech investments. Research by FinTech Global, Jan 2017
London leads growth in RegTech investments. Research by FinTech Global, Jan 2017
 
360 Degrees of FinTech (R)evolution
360 Degrees of FinTech (R)evolution360 Degrees of FinTech (R)evolution
360 Degrees of FinTech (R)evolution
 
The TransferWise Revolution - FinTech and the customer revolution in banking
The TransferWise Revolution - FinTech and the customer revolution in bankingThe TransferWise Revolution - FinTech and the customer revolution in banking
The TransferWise Revolution - FinTech and the customer revolution in banking
 

Semelhante a Artificial Intelligence for Mobile Visual Search

Internet Protocol Spoofing Presentation PPT
Internet Protocol Spoofing Presentation PPTInternet Protocol Spoofing Presentation PPT
Internet Protocol Spoofing Presentation PPT
manushreebhaskar
 

Semelhante a Artificial Intelligence for Mobile Visual Search (20)

The rise of VR AR - Lessons from the Pokemon`go
The rise of VR AR - Lessons from the Pokemon`goThe rise of VR AR - Lessons from the Pokemon`go
The rise of VR AR - Lessons from the Pokemon`go
 
Dissecting the dangers of deepfakes and their impact on reputation Generative...
Dissecting the dangers of deepfakes and their impact on reputation Generative...Dissecting the dangers of deepfakes and their impact on reputation Generative...
Dissecting the dangers of deepfakes and their impact on reputation Generative...
 
Creating Virtual Reality Training Using Interactive 360 Videos and Images wit...
Creating Virtual Reality Training Using Interactive 360 Videos and Images wit...Creating Virtual Reality Training Using Interactive 360 Videos and Images wit...
Creating Virtual Reality Training Using Interactive 360 Videos and Images wit...
 
Wearable Computing 시대, 플랫폼으로서의 AR(증강현실)의 비전
Wearable Computing 시대, 플랫폼으로서의 AR(증강현실)의 비전Wearable Computing 시대, 플랫폼으로서의 AR(증강현실)의 비전
Wearable Computing 시대, 플랫폼으로서의 AR(증강현실)의 비전
 
Mappatura delle eccellenze: Loop Ai Labs
Mappatura delle eccellenze: Loop Ai LabsMappatura delle eccellenze: Loop Ai Labs
Mappatura delle eccellenze: Loop Ai Labs
 
Augmented reality
Augmented realityAugmented reality
Augmented reality
 
Notes from Coursera Deep Learning courses by Andrew Ng
Notes from Coursera Deep Learning courses by Andrew NgNotes from Coursera Deep Learning courses by Andrew Ng
Notes from Coursera Deep Learning courses by Andrew Ng
 
How a library_visit_differentiates_your_brand_sxsw2105_jenessacarder_sapeient
How a library_visit_differentiates_your_brand_sxsw2105_jenessacarder_sapeientHow a library_visit_differentiates_your_brand_sxsw2105_jenessacarder_sapeient
How a library_visit_differentiates_your_brand_sxsw2105_jenessacarder_sapeient
 
CIS13: Is Identity the Answer to the Great Question of Life, the Universe, an...
CIS13: Is Identity the Answer to the Great Question of Life, the Universe, an...CIS13: Is Identity the Answer to the Great Question of Life, the Universe, an...
CIS13: Is Identity the Answer to the Great Question of Life, the Universe, an...
 
Data Ops:從實驗室走進生產線, 談談怎麼和資料科學家合作
Data Ops:從實驗室走進生產線, 談談怎麼和資料科學家合作Data Ops:從實驗室走進生產線, 談談怎麼和資料科學家合作
Data Ops:從實驗室走進生產線, 談談怎麼和資料科學家合作
 
KRDL April 2001
KRDL April 2001KRDL April 2001
KRDL April 2001
 
Closing the Gap: Bringing a Consumer-Like Experience to the Digital Workplace
Closing the Gap: Bringing a Consumer-Like Experience to the Digital WorkplaceClosing the Gap: Bringing a Consumer-Like Experience to the Digital Workplace
Closing the Gap: Bringing a Consumer-Like Experience to the Digital Workplace
 
Virtual reality(vr) assignment
Virtual reality(vr) assignmentVirtual reality(vr) assignment
Virtual reality(vr) assignment
 
Proceedings of INDEST 2014
Proceedings of INDEST 2014Proceedings of INDEST 2014
Proceedings of INDEST 2014
 
Trusted Data Ecosystems(信頼できるデータエコシステム):アイデンティティに価値を見出す
Trusted Data Ecosystems(信頼できるデータエコシステム):アイデンティティに価値を見出すTrusted Data Ecosystems(信頼できるデータエコシステム):アイデンティティに価値を見出す
Trusted Data Ecosystems(信頼できるデータエコシステム):アイデンティティに価値を見出す
 
Session 4: The Case for VR: Immersive Learning in the Workplace (Shelley Osbo...
Session 4: The Case for VR: Immersive Learning in the Workplace (Shelley Osbo...Session 4: The Case for VR: Immersive Learning in the Workplace (Shelley Osbo...
Session 4: The Case for VR: Immersive Learning in the Workplace (Shelley Osbo...
 
Digital spectacle by using cloud computing
Digital spectacle by using cloud computingDigital spectacle by using cloud computing
Digital spectacle by using cloud computing
 
Internet Protocol Spoofing Presentation PPT
Internet Protocol Spoofing Presentation PPTInternet Protocol Spoofing Presentation PPT
Internet Protocol Spoofing Presentation PPT
 
Artificial-Intelligence trends and its applications
Artificial-Intelligence trends and its applicationsArtificial-Intelligence trends and its applications
Artificial-Intelligence trends and its applications
 
Self-Sovereign Identity: Lightening Talk at RightsCon
Self-Sovereign Identity: Lightening Talk at RightsCon Self-Sovereign Identity: Lightening Talk at RightsCon
Self-Sovereign Identity: Lightening Talk at RightsCon
 

Mais de Swrve_Inc

Mais de Swrve_Inc (17)

Acquiring and Engaging Customers with Your Mobile App
Acquiring and Engaging Customers with Your Mobile AppAcquiring and Engaging Customers with Your Mobile App
Acquiring and Engaging Customers with Your Mobile App
 
Treat Your App Like It's Your Business
Treat Your App Like It's Your BusinessTreat Your App Like It's Your Business
Treat Your App Like It's Your Business
 
Mobile Commerce in the Age of Connected Devices
Mobile Commerce in the Age of Connected DevicesMobile Commerce in the Age of Connected Devices
Mobile Commerce in the Age of Connected Devices
 
Mobile Moments NYC 2016
Mobile Moments NYC 2016Mobile Moments NYC 2016
Mobile Moments NYC 2016
 
Resuscitating Your Omni-Channel Though Mobile
Resuscitating Your Omni-Channel Though MobileResuscitating Your Omni-Channel Though Mobile
Resuscitating Your Omni-Channel Though Mobile
 
Mobile Moments 2016 - Chris Babayode | Mobile Marketing Association
Mobile Moments 2016 - Chris Babayode | Mobile Marketing AssociationMobile Moments 2016 - Chris Babayode | Mobile Marketing Association
Mobile Moments 2016 - Chris Babayode | Mobile Marketing Association
 
Mobile Moments 2016 - Gearóid O'Rourke | Helpling
Mobile Moments 2016 - Gearóid O'Rourke | HelplingMobile Moments 2016 - Gearóid O'Rourke | Helpling
Mobile Moments 2016 - Gearóid O'Rourke | Helpling
 
Mobile Moments 2016 - Adam Warbuton | Travelex
Mobile Moments 2016 - Adam Warbuton | TravelexMobile Moments 2016 - Adam Warbuton | Travelex
Mobile Moments 2016 - Adam Warbuton | Travelex
 
Mobile Moments 2016 - Ben Phillips - Media Com
Mobile Moments 2016 -  Ben Phillips - Media ComMobile Moments 2016 -  Ben Phillips - Media Com
Mobile Moments 2016 - Ben Phillips - Media Com
 
Mobile Moments 2016 - Martin McKenna | iZettle
Mobile Moments 2016 -  Martin McKenna | iZettleMobile Moments 2016 -  Martin McKenna | iZettle
Mobile Moments 2016 - Martin McKenna | iZettle
 
Mobile is the New Retail
Mobile is the New Retail      Mobile is the New Retail
Mobile is the New Retail
 
Mobile And The Media 5 Ways To Succeed In 2016
Mobile And The Media 5 Ways To Succeed In 2016Mobile And The Media 5 Ways To Succeed In 2016
Mobile And The Media 5 Ways To Succeed In 2016
 
Banking On Mobile - Getting Ready for 2016
Banking On Mobile -  Getting Ready for 2016Banking On Mobile -  Getting Ready for 2016
Banking On Mobile - Getting Ready for 2016
 
Mobile: Shaping the First-screen Customer Experience
Mobile: Shaping the First-screen Customer ExperienceMobile: Shaping the First-screen Customer Experience
Mobile: Shaping the First-screen Customer Experience
 
Optimizing first time user experiences
Optimizing first time user experiencesOptimizing first time user experiences
Optimizing first time user experiences
 
Tracking, Testing, Targeting: Getting Consumers To Monetize On Mobile
Tracking, Testing, Targeting: Getting Consumers To Monetize On MobileTracking, Testing, Targeting: Getting Consumers To Monetize On Mobile
Tracking, Testing, Targeting: Getting Consumers To Monetize On Mobile
 
5 Killer Mistakes in Mobile - 3xe Digital Conference 2015.
5 Killer Mistakes in Mobile - 3xe Digital Conference 2015.5 Killer Mistakes in Mobile - 3xe Digital Conference 2015.
5 Killer Mistakes in Mobile - 3xe Digital Conference 2015.
 

Último

Obat Penggugur Kandungan Di Apotik Kimia Farma (087776558899)
Obat Penggugur Kandungan Di Apotik Kimia Farma (087776558899)Obat Penggugur Kandungan Di Apotik Kimia Farma (087776558899)
Obat Penggugur Kandungan Di Apotik Kimia Farma (087776558899)
Cara Menggugurkan Kandungan 087776558899
 

Último (6)

FULL ENJOY - 9999218229 Call Girls in {Mahipalpur}| Delhi NCR
FULL ENJOY - 9999218229 Call Girls in {Mahipalpur}| Delhi NCRFULL ENJOY - 9999218229 Call Girls in {Mahipalpur}| Delhi NCR
FULL ENJOY - 9999218229 Call Girls in {Mahipalpur}| Delhi NCR
 
BDSM⚡Call Girls in Sector 71 Noida Escorts >༒8448380779 Escort Service
BDSM⚡Call Girls in Sector 71 Noida Escorts >༒8448380779 Escort ServiceBDSM⚡Call Girls in Sector 71 Noida Escorts >༒8448380779 Escort Service
BDSM⚡Call Girls in Sector 71 Noida Escorts >༒8448380779 Escort Service
 
Leading Mobile App Development Companies in India (2).pdf
Leading Mobile App Development Companies in India (2).pdfLeading Mobile App Development Companies in India (2).pdf
Leading Mobile App Development Companies in India (2).pdf
 
Obat Penggugur Kandungan Di Apotik Kimia Farma (087776558899)
Obat Penggugur Kandungan Di Apotik Kimia Farma (087776558899)Obat Penggugur Kandungan Di Apotik Kimia Farma (087776558899)
Obat Penggugur Kandungan Di Apotik Kimia Farma (087776558899)
 
9999266834 Call Girls In Noida Sector 52 (Delhi) Call Girl Service
9999266834 Call Girls In Noida Sector 52 (Delhi) Call Girl Service9999266834 Call Girls In Noida Sector 52 (Delhi) Call Girl Service
9999266834 Call Girls In Noida Sector 52 (Delhi) Call Girl Service
 
Powerful Love Spells in Arkansas, AR (310) 882-6330 Bring Back Lost Lover
Powerful Love Spells in Arkansas, AR (310) 882-6330 Bring Back Lost LoverPowerful Love Spells in Arkansas, AR (310) 882-6330 Bring Back Lost Lover
Powerful Love Spells in Arkansas, AR (310) 882-6330 Bring Back Lost Lover
 

Artificial Intelligence for Mobile Visual Search

Notas do Editor

  1. Hello - my name is Catherine Ulrich and I am the Chief Product Officer at Shutterstock. Shutterstock is a technology company that enables marketers, designers, businesses and creative professionals to find and license royalty free images, video and music for use on websites, ad campaigns, marketing materials and movies. More than 100,000 photographers, artists, videographers and musicians upload and license their work via Shutterstock. Uploading more than 800,000 images/video/music clips a week. 1.5million customers. I’m going to share some insights from leveraging artificial intelligence for visual search - I’ll focus on How we taught the technology How we leveraged that for mobile
  2. The biggest challenge, faced by people searching for images. I found this image on the internet and I’d really like to use it in a marketing campaign I’m working on. Since I don’t own the rights to this image, I need to purchase a licensed version of it so I’ll try to describe this image to find something similar on Shutterstock. I might use the words stained, glass, window, church, angle.
  3. As you can see it is challenging. More keywords do not necessarily help improve the accuracy, and in fact about 65% of searches are just 1 or 2 keywords on Shutterstock - it’s not a lot for us to go off of. The results coming up are all relevant to the keywords, but they don’t look like the image I want. I’m not sure the words exist to describe what it is I like about this image to put into the search box. 1. Language accuracy challenge 2. Articulating how a image looks 3. Relevancy/popular based on downloads
  4. To help customers with this issue we created a Computer Vision team to leverage artificial intelligence for improved search functionality. In 2012, there was a huge advancement in computer vision with the publication of the Krachevsky paper on imagenet classification. This is the first model our team used to build an initial service that was poor and slow, but showed promise.
  5. A Neural Network is a system modeled after the human brain, and the way it works is that you feed in a large number of images with labels, such as cat pictures, and the system literally “learns” what a cat looks like. The more training pictures you give it, the more it can learn. The more the system understands what is in an image, the better it is able to compute similarity. And that’s great for us, because at Shutterstock we have a LOT of high-quality data. We sell about 5 images a second, or about 400,000 images a day. Each of these are preceded by searches, clicks, detail page views, etc, which sums up to a LOT of training data. Over the past year they iterated on making it smarter and faster, until it reached the quality level we were happy to share publicly. I’ll show you the process of refining this and how we continue to improve this model.
  6. The computer vision team within Shutterstock harnessed this data and were given the explicit mission of advancing our understanding of images to power next generation search experiences. Here’s how it began to evolve over time.
  7. Constantly refining, almost like two dials - trying to balance the pixel data that we’re feeding the model.
  8. Picking up the colors. But only one face. No geometric pattern - but understanding the abstract.
  9. Hyper focused on the facial elements and colors, artistic style
  10. Too far on the geometric side of the scale - meant the facial features are lost.
  11. This is the model we launched with and continue to refine with more data every day as customers search and download. Both images are really some of the more complicated images we fed the model - in terms of composition. This is to show how far its come.
  12. Now with reverse image for desktop launched in March - these are the results.
  13. If you remember from the beginning, this is the image I wanted to describe and these were the results I got. Launched on web in March.
  14. Yesterday we launched reverse image search for mobile. This is revolutionary for our customers to be able to search for images on the go. We’re the only stock photo provider to have launched the search/download computer vision technology on mobile. Step by step demo -
  15. As we hoped, reverse image search is benefitting customers in non-English speaking countries most, as we can see by the share of reverse image downloads vs. traditional text search downloads by country: For our on-demand customers, this is vastly outperforming text search. Continue to improve and evolve as the collection grows. Different asset types and Android later in the summer. Transcending language.