Soumettre la recherche
Mettre en ligne
Making advertising personal, 4th NL Recommenders Meetup
•
Télécharger en tant que PPTX, PDF
•
1 j'aime
•
886 vues
O
Olivier Koch
Suivre
We present the Criteo recommendation stack
Lire moins
Lire la suite
Technologie
Affichage du diaporama
Signaler
Partager
Affichage du diaporama
Signaler
Partager
1 sur 15
Télécharger maintenant
Recommandé
Criteo TektosData Meetup
Criteo TektosData Meetup
Olivier Koch
New challenges for scalable machine learning in online advertising
New challenges for scalable machine learning in online advertising
Olivier Koch
New machine learning challenges at Criteo
New machine learning challenges at Criteo
Olivier Koch
Damien Lefortier, Senior Machine Learning Engineer and Tech Lead in the Predi...
Damien Lefortier, Senior Machine Learning Engineer and Tech Lead in the Predi...
MLconf
Machine learning at Criteo - Paris Datageeks
Machine learning at Criteo - Paris Datageeks
Nicolas Le Roux
EN - Criteo - BD Deck -July 2014 - rebrand
EN - Criteo - BD Deck -July 2014 - rebrand
Djilali Zitouni
RecSys 2015: Large-scale real-time product recommendation at Criteo
RecSys 2015: Large-scale real-time product recommendation at Criteo
Romain Lerallut
Amanda Casari, Senior Data Scientist, Concur at MLconf SEA - 5/20/16
Amanda Casari, Senior Data Scientist, Concur at MLconf SEA - 5/20/16
MLconf
Recommandé
Criteo TektosData Meetup
Criteo TektosData Meetup
Olivier Koch
New challenges for scalable machine learning in online advertising
New challenges for scalable machine learning in online advertising
Olivier Koch
New machine learning challenges at Criteo
New machine learning challenges at Criteo
Olivier Koch
Damien Lefortier, Senior Machine Learning Engineer and Tech Lead in the Predi...
Damien Lefortier, Senior Machine Learning Engineer and Tech Lead in the Predi...
MLconf
Machine learning at Criteo - Paris Datageeks
Machine learning at Criteo - Paris Datageeks
Nicolas Le Roux
EN - Criteo - BD Deck -July 2014 - rebrand
EN - Criteo - BD Deck -July 2014 - rebrand
Djilali Zitouni
RecSys 2015: Large-scale real-time product recommendation at Criteo
RecSys 2015: Large-scale real-time product recommendation at Criteo
Romain Lerallut
Amanda Casari, Senior Data Scientist, Concur at MLconf SEA - 5/20/16
Amanda Casari, Senior Data Scientist, Concur at MLconf SEA - 5/20/16
MLconf
criteo-performance-advertising-playbook-2015
criteo-performance-advertising-playbook-2015
Carolyn Bednarz
Introduction Criteo - 2.0
Introduction Criteo - 2.0
Scott Turecek
Simon Dollé_Large-scale Real-time recommendation at Criteo
Simon Dollé_Large-scale Real-time recommendation at Criteo
Dataconomy Media
Criteo Infrastructure (Platform) Meetup
Criteo Infrastructure (Platform) Meetup
Ibrahim Abubakari
Criteo Couchbase live 2015
Criteo Couchbase live 2015
Nicolasgmail.com Helleringer
Your Future With Content Manager OnDemand
Your Future With Content Manager OnDemand
Zia Consulting
Aws community day keynote
Aws community day keynote
James Samuel
Ad Server Solutions - ad server ad exchange
Ad Server Solutions - ad server ad exchange
Ad Server Solutions
3 Minute Introduction
3 Minute Introduction
Julian Tol
Optimize your Cloud Purchase Strategy
Optimize your Cloud Purchase Strategy
Jan Thielscher
Google’s strategy to gain an edge on the Cloud
Google’s strategy to gain an edge on the Cloud
RohitSingh1837
Successful IoT projects - a few lessons
Successful IoT projects - a few lessons
Jan Thielscher
Online Ad Serving
Online Ad Serving
Neha Gupta
Criteo's Ad Week 2012 presentation - Big Data and the Value of Clickers
Criteo's Ad Week 2012 presentation - Big Data and the Value of Clickers
Criteo
Axonite Campaign Automation Infrastructure for HasOffers
Axonite Campaign Automation Infrastructure for HasOffers
Yuval Shefler
CrowdCast Pitch
CrowdCast Pitch
Shahrukh Ghazali
C# development workflow @ criteo
C# development workflow @ criteo
Ibrahim Abubakari
RecsysFR: Criteo presentation
RecsysFR: Criteo presentation
recsysfr
Criteo. Reach people, not devices!
Criteo. Reach people, not devices!
HybridRussia
Why reinvent the wheel at Criteo?
Why reinvent the wheel at Criteo?
Criteolabs
Couchbase live 2016
Couchbase live 2016
Pierre Mavro
Response prediction for display advertising - WSDM 2014
Response prediction for display advertising - WSDM 2014
Olivier Chapelle
Contenu connexe
Tendances
criteo-performance-advertising-playbook-2015
criteo-performance-advertising-playbook-2015
Carolyn Bednarz
Introduction Criteo - 2.0
Introduction Criteo - 2.0
Scott Turecek
Simon Dollé_Large-scale Real-time recommendation at Criteo
Simon Dollé_Large-scale Real-time recommendation at Criteo
Dataconomy Media
Criteo Infrastructure (Platform) Meetup
Criteo Infrastructure (Platform) Meetup
Ibrahim Abubakari
Criteo Couchbase live 2015
Criteo Couchbase live 2015
Nicolasgmail.com Helleringer
Your Future With Content Manager OnDemand
Your Future With Content Manager OnDemand
Zia Consulting
Aws community day keynote
Aws community day keynote
James Samuel
Ad Server Solutions - ad server ad exchange
Ad Server Solutions - ad server ad exchange
Ad Server Solutions
3 Minute Introduction
3 Minute Introduction
Julian Tol
Optimize your Cloud Purchase Strategy
Optimize your Cloud Purchase Strategy
Jan Thielscher
Google’s strategy to gain an edge on the Cloud
Google’s strategy to gain an edge on the Cloud
RohitSingh1837
Successful IoT projects - a few lessons
Successful IoT projects - a few lessons
Jan Thielscher
Online Ad Serving
Online Ad Serving
Neha Gupta
Criteo's Ad Week 2012 presentation - Big Data and the Value of Clickers
Criteo's Ad Week 2012 presentation - Big Data and the Value of Clickers
Criteo
Axonite Campaign Automation Infrastructure for HasOffers
Axonite Campaign Automation Infrastructure for HasOffers
Yuval Shefler
CrowdCast Pitch
CrowdCast Pitch
Shahrukh Ghazali
Tendances
(16)
criteo-performance-advertising-playbook-2015
criteo-performance-advertising-playbook-2015
Introduction Criteo - 2.0
Introduction Criteo - 2.0
Simon Dollé_Large-scale Real-time recommendation at Criteo
Simon Dollé_Large-scale Real-time recommendation at Criteo
Criteo Infrastructure (Platform) Meetup
Criteo Infrastructure (Platform) Meetup
Criteo Couchbase live 2015
Criteo Couchbase live 2015
Your Future With Content Manager OnDemand
Your Future With Content Manager OnDemand
Aws community day keynote
Aws community day keynote
Ad Server Solutions - ad server ad exchange
Ad Server Solutions - ad server ad exchange
3 Minute Introduction
3 Minute Introduction
Optimize your Cloud Purchase Strategy
Optimize your Cloud Purchase Strategy
Google’s strategy to gain an edge on the Cloud
Google’s strategy to gain an edge on the Cloud
Successful IoT projects - a few lessons
Successful IoT projects - a few lessons
Online Ad Serving
Online Ad Serving
Criteo's Ad Week 2012 presentation - Big Data and the Value of Clickers
Criteo's Ad Week 2012 presentation - Big Data and the Value of Clickers
Axonite Campaign Automation Infrastructure for HasOffers
Axonite Campaign Automation Infrastructure for HasOffers
CrowdCast Pitch
CrowdCast Pitch
En vedette
C# development workflow @ criteo
C# development workflow @ criteo
Ibrahim Abubakari
RecsysFR: Criteo presentation
RecsysFR: Criteo presentation
recsysfr
Criteo. Reach people, not devices!
Criteo. Reach people, not devices!
HybridRussia
Why reinvent the wheel at Criteo?
Why reinvent the wheel at Criteo?
Criteolabs
Couchbase live 2016
Couchbase live 2016
Pierre Mavro
Response prediction for display advertising - WSDM 2014
Response prediction for display advertising - WSDM 2014
Olivier Chapelle
Infographic: How Professional Services Automation can transform your Company!
Infographic: How Professional Services Automation can transform your Company!
Changepoint
Saintjo Two AV4
Saintjo Two AV4
Bernard Alexandre
Bizarre... vous avez dit bizarre - Paris Monitoring meetup #2
Bizarre... vous avez dit bizarre - Paris Monitoring meetup #2
Paris Monitoring
Etude Criteo - la valeur réelle des internautes qui cliquent - 26 Juin 2012
Etude Criteo - la valeur réelle des internautes qui cliquent - 26 Juin 2012
Romain Fonnier
Morning with MongoDB Paris 2012 - Cas d'usages courant en entreprise. Présent...
Morning with MongoDB Paris 2012 - Cas d'usages courant en entreprise. Présent...
MongoDB
AWS Summit Paris - Track 1 - Session 2 - Designez vos architectures pour plus...
AWS Summit Paris - Track 1 - Session 2 - Designez vos architectures pour plus...
Amazon Web Services
Hadoop summit-ams-2014-04-03
Hadoop summit-ams-2014-04-03
SDanzanvilliersCriteo
TechFeedというテクノロジーキュレーションサービスを作った話
TechFeedというテクノロジーキュレーションサービスを作った話
yoshikawa_t
Using Neural Networks to predict user ratings
Using Neural Networks to predict user ratings
recsysfr
Dictionary Learning for Massive Matrix Factorization
Dictionary Learning for Massive Matrix Factorization
recsysfr
Meta-Prod2Vec: Simple Product Embeddings with Side-Information
Meta-Prod2Vec: Simple Product Embeddings with Side-Information
recsysfr
CONTENT2VEC: a Joint Architecture to use Product Image and Text for the task ...
CONTENT2VEC: a Joint Architecture to use Product Image and Text for the task ...
recsysfr
En vedette
(18)
C# development workflow @ criteo
C# development workflow @ criteo
RecsysFR: Criteo presentation
RecsysFR: Criteo presentation
Criteo. Reach people, not devices!
Criteo. Reach people, not devices!
Why reinvent the wheel at Criteo?
Why reinvent the wheel at Criteo?
Couchbase live 2016
Couchbase live 2016
Response prediction for display advertising - WSDM 2014
Response prediction for display advertising - WSDM 2014
Infographic: How Professional Services Automation can transform your Company!
Infographic: How Professional Services Automation can transform your Company!
Saintjo Two AV4
Saintjo Two AV4
Bizarre... vous avez dit bizarre - Paris Monitoring meetup #2
Bizarre... vous avez dit bizarre - Paris Monitoring meetup #2
Etude Criteo - la valeur réelle des internautes qui cliquent - 26 Juin 2012
Etude Criteo - la valeur réelle des internautes qui cliquent - 26 Juin 2012
Morning with MongoDB Paris 2012 - Cas d'usages courant en entreprise. Présent...
Morning with MongoDB Paris 2012 - Cas d'usages courant en entreprise. Présent...
AWS Summit Paris - Track 1 - Session 2 - Designez vos architectures pour plus...
AWS Summit Paris - Track 1 - Session 2 - Designez vos architectures pour plus...
Hadoop summit-ams-2014-04-03
Hadoop summit-ams-2014-04-03
TechFeedというテクノロジーキュレーションサービスを作った話
TechFeedというテクノロジーキュレーションサービスを作った話
Using Neural Networks to predict user ratings
Using Neural Networks to predict user ratings
Dictionary Learning for Massive Matrix Factorization
Dictionary Learning for Massive Matrix Factorization
Meta-Prod2Vec: Simple Product Embeddings with Side-Information
Meta-Prod2Vec: Simple Product Embeddings with Side-Information
CONTENT2VEC: a Joint Architecture to use Product Image and Text for the task ...
CONTENT2VEC: a Joint Architecture to use Product Image and Text for the task ...
Similaire à Making advertising personal, 4th NL Recommenders Meetup
Recommendation at scale
Recommendation at scale
simondolle
Open Sourcing GemFire - Apache Geode
Open Sourcing GemFire - Apache Geode
Apache Geode
An Introduction to Apache Geode (incubating)
An Introduction to Apache Geode (incubating)
Anthony Baker
GraphTalk Berlin - Einführung in Graphdatenbanken
GraphTalk Berlin - Einführung in Graphdatenbanken
Neo4j
Webinar: 2 Billion Data Points Each Day
Webinar: 2 Billion Data Points Each Day
DataStax
The New Model
The New Model
David Kaiser
10 Lessons Learned from Meeting with 150 Banks Across the Globe
10 Lessons Learned from Meeting with 150 Banks Across the Globe
DataWorks Summit
OC Big Data Monthly Meetup #6 - Session 1 - IBM
OC Big Data Monthly Meetup #6 - Session 1 - IBM
Big Data Joe™ Rossi
SD Big Data Monthly Meetup #4 - Session 1 - IBM
SD Big Data Monthly Meetup #4 - Session 1 - IBM
Big Data Joe™ Rossi
RTBkit Introduction & Best Practices
RTBkit Introduction & Best Practices
Datacratic
Hybris @ Neev
Hybris @ Neev
Neev Technologies
Connecta Event: Big Query och dataanalys med Google Cloud Platform
Connecta Event: Big Query och dataanalys med Google Cloud Platform
ConnectaDigital
The Business Justification for APM
The Business Justification for APM
Jonah Kowall
Conclusion Connect state of IoT 2019 Review io t solutions world congress 2019
Conclusion Connect state of IoT 2019 Review io t solutions world congress 2019
Conclusion Connect enabling industry 4.0 with IoT
Dashlane Mission Teams
Dashlane Mission Teams
Dashlane
Lean Startup: Reduce 40% go-to-market time & cost on your next product launch
Lean Startup: Reduce 40% go-to-market time & cost on your next product launch
People10 Technosoft Private Limited
AI as a Service, Build Shared AI Service Platforms Based on Deep Learning Tec...
AI as a Service, Build Shared AI Service Platforms Based on Deep Learning Tec...
Databricks
Lean startup 2019
Lean startup 2019
John J. H. Oh
Self-Service Analytics with Guard Rails
Self-Service Analytics with Guard Rails
Denodo
Containers and VMs and Clouds: Oh My. by Mike Coleman
Containers and VMs and Clouds: Oh My. by Mike Coleman
Docker, Inc.
Similaire à Making advertising personal, 4th NL Recommenders Meetup
(20)
Recommendation at scale
Recommendation at scale
Open Sourcing GemFire - Apache Geode
Open Sourcing GemFire - Apache Geode
An Introduction to Apache Geode (incubating)
An Introduction to Apache Geode (incubating)
GraphTalk Berlin - Einführung in Graphdatenbanken
GraphTalk Berlin - Einführung in Graphdatenbanken
Webinar: 2 Billion Data Points Each Day
Webinar: 2 Billion Data Points Each Day
The New Model
The New Model
10 Lessons Learned from Meeting with 150 Banks Across the Globe
10 Lessons Learned from Meeting with 150 Banks Across the Globe
OC Big Data Monthly Meetup #6 - Session 1 - IBM
OC Big Data Monthly Meetup #6 - Session 1 - IBM
SD Big Data Monthly Meetup #4 - Session 1 - IBM
SD Big Data Monthly Meetup #4 - Session 1 - IBM
RTBkit Introduction & Best Practices
RTBkit Introduction & Best Practices
Hybris @ Neev
Hybris @ Neev
Connecta Event: Big Query och dataanalys med Google Cloud Platform
Connecta Event: Big Query och dataanalys med Google Cloud Platform
The Business Justification for APM
The Business Justification for APM
Conclusion Connect state of IoT 2019 Review io t solutions world congress 2019
Conclusion Connect state of IoT 2019 Review io t solutions world congress 2019
Dashlane Mission Teams
Dashlane Mission Teams
Lean Startup: Reduce 40% go-to-market time & cost on your next product launch
Lean Startup: Reduce 40% go-to-market time & cost on your next product launch
AI as a Service, Build Shared AI Service Platforms Based on Deep Learning Tec...
AI as a Service, Build Shared AI Service Platforms Based on Deep Learning Tec...
Lean startup 2019
Lean startup 2019
Self-Service Analytics with Guard Rails
Self-Service Analytics with Guard Rails
Containers and VMs and Clouds: Oh My. by Mike Coleman
Containers and VMs and Clouds: Oh My. by Mike Coleman
Dernier
Automating Google Workspace (GWS) & more with Apps Script
Automating Google Workspace (GWS) & more with Apps Script
wesley chun
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
Joaquim Jorge
Histor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slide
vu2urc
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Miguel Araújo
Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024
The Digital Insurer
A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?
Igalia
AWS Community Day CPH - Three problems of Terraform
AWS Community Day CPH - Three problems of Terraform
Andrey Devyatkin
Data Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt Robison
Anna Loughnan Colquhoun
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
The Digital Insurer
🐬 The future of MySQL is Postgres 🐘
🐬 The future of MySQL is Postgres 🐘
RTylerCroy
Real Time Object Detection Using Open CV
Real Time Object Detection Using Open CV
Khem
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
ThousandEyes
Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...
apidays
presentation ICT roal in 21st century education
presentation ICT roal in 21st century education
jfdjdjcjdnsjd
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed texts
Maria Levchenko
What Are The Drone Anti-jamming Systems Technology?
What Are The Drone Anti-jamming Systems Technology?
Antenna Manufacturer Coco
Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024
The Digital Insurer
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
The Digital Insurer
Advantages of Hiring UIUX Design Service Providers for Your Business
Advantages of Hiring UIUX Design Service Providers for Your Business
Pixlogix Infotech
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc
Dernier
(20)
Automating Google Workspace (GWS) & more with Apps Script
Automating Google Workspace (GWS) & more with Apps Script
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
Histor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slide
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024
A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?
AWS Community Day CPH - Three problems of Terraform
AWS Community Day CPH - Three problems of Terraform
Data Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt Robison
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
🐬 The future of MySQL is Postgres 🐘
🐬 The future of MySQL is Postgres 🐘
Real Time Object Detection Using Open CV
Real Time Object Detection Using Open CV
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...
presentation ICT roal in 21st century education
presentation ICT roal in 21st century education
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed texts
What Are The Drone Anti-jamming Systems Technology?
What Are The Drone Anti-jamming Systems Technology?
Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
Advantages of Hiring UIUX Design Service Providers for Your Business
Advantages of Hiring UIUX Design Service Providers for Your Business
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
Making advertising personal, 4th NL Recommenders Meetup
1.
Copyright © 2014
Criteo Making Advertising Personal Large Scale Real-Time Product Recommendation at Criteo Olivier Koch & Romain Lerallut May 19th, 2015
2.
Copyright © 2014
Criteo Performance Advertising We buy • Inventory ! (ad spaces) • Billions of times a day • All over the Internet • For 95% of the population Where is the need for tech ? We sell • Clicks ! • (that convert) • (that convert a lot) We take the risk You pay only for what you get
3.
Copyright © 2013.
Confidential International infrastructure Key figures Traffic 550 k HTTP requests / sec (peak activity) 23000 impressions /sec (peak activity) 180 k requests / sec on RTB (average) Less than 10 ms to process an RTB request 3 Figures of May2013 Physical infrastructure 6 Data centers on 3 continents operated and conceived in- house ~ 12000 servers, largest Hadoop cluster in Europe Availability / Uptime >99.95% More than 20 PB of storage Big Data
4.
Copyright © 2012.
Confidential 4 Arbitrage •Should we bid? •At which price? Recommendation •Which products should we display? Graphical optimization •Big image vs small image •Background color, ... Prediction •Generic prediction engine •Specific models trained on TBs of data Global Engine Architecture
5.
Copyright © 2014
Criteo Building ads with personalized content in realtime Recommendation for Advertising 2.5 billion products ~ 8ms response time
6.
Copyright © 2014
Criteo Data Sources • Catalog data • Feed provided by the merchants • User behavior data • Large scale intent data • All visits to merchant websites • Page views, basket, sales events • Ad display data • Displayed and clicked ads 6
7.
Copyright © 2012.
Confidential 7 Recommendation execution flow Candidates Generation • Get candidates from all sources using user historical products and bestofs Candidates Aggregation • Remove duplicates • Aggregate features Preselection • Call degraded prediction to decrease number of candidates Scoring • Call full prediction model to score each candidate Winner Selector • Select N products on score, randomization Glup Logging • Log recommended products, prediction variables
8.
Copyright © 2014
Criteo Issue #1: Retrieving user-specific products • The need: storing [user → interesting products] vectors • Difficult to store and retrieve at scale (900+ mln users) • Hard to keep up to date • Using seen products as a proxy • Store the [user → viewed products] vectors • Easier to maintain • Store [viewed product → interesting products] vectors • Based on aggregated user behavior data • Computable offline • Final ranking by a ML model 8
9.
Copyright © 2014
Criteo Issue #2: Scoring products • Need to fuse data from several sources • Product-specific • User-specific • User-product interactions • Display-specific • 1st solution: regression model • Predict P(product click then sale) • Easy to evaluate • 2nd solution: ranking model • More appropriate for our needs • Still maximizing post-click sales • Can be evaluated only on multi-product banners 9
10.
Copyright © 2014
Criteo Issue #3: Picking the right products • Several questions: • User-product fatigue • Independent product choice assumption • Explore / exploit • Solution 1: Randomization in the banners • Keep independent products assumption • Separate optimization process to shuffle displayed items • Solution 2: A better scoring model • Score a full banner, not independent products • Store all product display counts 10
11.
Copyright © 2014
Criteo Issue #4: 8ms response time ! Tweaking for performance • CTR / Sales prediction takes ~40 µs per candidate using in-house library • 2-step prediction: • A fast pass to remove most of the candidates • A slow pass to score accurately the remaining candidates And the technical fine print: • All real-time code in C# • Async I/O for better efficiency • HAProxy to scale the front-end • Memcached to store all required data in memory 11
12.
Copyright © 2014
Criteo Upcoming challenges • Long(er)-term user profiles • More and better product information (images, semantic, NLP) • Vertical-optimized engine • Classifieds (catalog-free recommendation ?) • Travel • Instant-update of similarities • (because batch computation is soooo last year) 12
13.
Copyright © 2014
Criteo Fancy a try ? 13 On your own: With us ! http://www.criteo.com/careers/ • Our 1st public dataset is online: http://bit.ly/1vgw2XC • 4GB display and click data, Kaggle challenge in 2014 • NEW : 1TB dataset released a few weeks ago • Hosted on Microsoft Azure, just waiting for you
14.
Copyright © 2014
Criteo Questions?
15.
Copyright © 2014
Criteo Thank you ! o.koch@criteo.com r.lerallut@criteo.com
Télécharger maintenant