SlideShare a Scribd company logo
1 of 27
Challenges and Research for
Real-Time Recommendation
in a Dynamic Marketplace
Environment
Some unlimited consumption items
Limited but with stock under control
Unique items and stock out of control
@timotta
contacts
influenced by
recommendations
Largest online classified in Brazil
daily users
+7M +500k
new ads
published daily
+15%
Heavy ad publishing flow
X
Graph based collaborative filtering
Real-time graph updating
Adega
(PostgreSQL)
Sommelier
(API)
Lurker
(Tracker)
Stream
processor
contacts
Great result
adviews
+6% +4%
Collaborative filtering based on ad views
View
Contact
Concentration
Buyers
Sellers
:(
X
X
✓
:(
:(
The Idea:
A content-based
ranked by contact
probability
Random item-item balanced dataset
Ad Viewed
Features
Ad
Recommended
Features
Target:
Contacted
yes or no?
Features: Title and description
Title, description
and category
Doc2Vec
embedding
Features: Image
Image embedding
from ResNet's
penultimate layer
Features: Neighborhood
Neighborhood
latent factors generated by
logistic matrix factorization
Features: Price $$$
Price
Classification on a balanced dataset
accuracy
75%
Studying how to
compare both
methods offline . . .
Cannot predict online
due to high time
loading candidate ads
and calculating
probabilities
Real-time background prediction
Adega
(PostgreSQL)
Embedding
calculations
Real-time background prediction
Adega
(PostgreSQL)
Embedding
calculations
Probability
calculation
Real-time background prediction
Adega
(PostgreSQL)
Embedding
calculations
Probability
calculation
Reversed
Probability
calculation
Real-time background prediction
Adega
(PostgreSQL)
Sommelier
(API)
Embedding
calculations
Probability
calculation
Reversed
Probability
calculation
Future research
News Session-Based Recommendations
using Deep Neural Networks (Chamaleon)
Metadata Embeddings for User and Item
Cold-start Recommendations (lightFM)
Recommendation Squad at OLX
Filipe
Casal
Marcelo
Malta
Tiago
Motta
Leonardo
Wajnsztok
Thays
Macedo

More Related Content

Similar to Challenges of Real-Time Recommendations in a Dynamic Marketplace

RTB-Media Intro
RTB-Media IntroRTB-Media Intro
RTB-Media IntroRTB-Media
 
Qcon London 2017 - Architecture overhaul - Ad serving @ Spotify scale
Qcon London 2017 -  Architecture overhaul - Ad serving @ Spotify scaleQcon London 2017 -  Architecture overhaul - Ad serving @ Spotify scale
Qcon London 2017 - Architecture overhaul - Ad serving @ Spotify scaleKinshuk Mishra
 
SEMrush: Killer Features
SEMrush: Killer FeaturesSEMrush: Killer Features
SEMrush: Killer FeaturesTopOnSeek
 
Marketplace in motion - AdKDD keynote - 2020
Marketplace in motion - AdKDD keynote - 2020 Marketplace in motion - AdKDD keynote - 2020
Marketplace in motion - AdKDD keynote - 2020 Roelof van Zwol
 
Computational Marketing at Groupon - JCSSE 2017
Computational Marketing at Groupon - JCSSE 2017Computational Marketing at Groupon - JCSSE 2017
Computational Marketing at Groupon - JCSSE 2017Clovis Chapman
 
Flip myfunnel raab v2
Flip myfunnel raab v2Flip myfunnel raab v2
Flip myfunnel raab v2draab
 
Ad Extensions as a Lever for Powerful Text Ads
Ad Extensions as a Lever for Powerful Text AdsAd Extensions as a Lever for Powerful Text Ads
Ad Extensions as a Lever for Powerful Text AdsHanapin Marketing
 
Recsys2016 Tutorial by Xavier and Deepak
Recsys2016 Tutorial by Xavier and DeepakRecsys2016 Tutorial by Xavier and Deepak
Recsys2016 Tutorial by Xavier and DeepakDeepak Agarwal
 
TMKu: Campaign Infrastructure
TMKu: Campaign InfrastructureTMKu: Campaign Infrastructure
TMKu: Campaign InfrastructureThe Media Kitchen
 
Decentralized Advertising Marketplace
Decentralized Advertising MarketplaceDecentralized Advertising Marketplace
Decentralized Advertising MarketplaceNedko Nedkov
 
How to Segment Your Customers and Target Their Needs
How to Segment Your Customers and Target Their NeedsHow to Segment Your Customers and Target Their Needs
How to Segment Your Customers and Target Their NeedsWebtrends
 
Find it! Nail it! Boosting e-commerce search conversions with machine learnin...
Find it! Nail it!Boosting e-commerce search conversions with machine learnin...Find it! Nail it!Boosting e-commerce search conversions with machine learnin...
Find it! Nail it! Boosting e-commerce search conversions with machine learnin...Rakuten Group, Inc.
 
Google Ad Yield Management (2016 Feb) by Acqua Media
Google Ad Yield Management (2016 Feb) by Acqua MediaGoogle Ad Yield Management (2016 Feb) by Acqua Media
Google Ad Yield Management (2016 Feb) by Acqua Mediarally1275
 
Google Analytics Premium for Better Data-Driven Decisions With Swapnil Sinha
Google Analytics Premium for Better Data-Driven Decisions With Swapnil SinhaGoogle Analytics Premium for Better Data-Driven Decisions With Swapnil Sinha
Google Analytics Premium for Better Data-Driven Decisions With Swapnil SinhaTatvic Analytics
 
Intro to Programmatic Advertising with Matt Prohaska from Prohaska Consulting
Intro to Programmatic Advertising with Matt Prohaska from Prohaska ConsultingIntro to Programmatic Advertising with Matt Prohaska from Prohaska Consulting
Intro to Programmatic Advertising with Matt Prohaska from Prohaska ConsultingStukent Inc.
 
Cross channel attribution overview feb 2010
Cross channel attribution overview feb 2010Cross channel attribution overview feb 2010
Cross channel attribution overview feb 2010xplusone
 
Brian Clifton Keynote at Analytics That Excite
Brian Clifton Keynote at Analytics That ExciteBrian Clifton Keynote at Analytics That Excite
Brian Clifton Keynote at Analytics That ExciteInfoTrust LLC
 

Similar to Challenges of Real-Time Recommendations in a Dynamic Marketplace (20)

RTB-Media Intro
RTB-Media IntroRTB-Media Intro
RTB-Media Intro
 
Qcon London 2017 - Architecture overhaul - Ad serving @ Spotify scale
Qcon London 2017 -  Architecture overhaul - Ad serving @ Spotify scaleQcon London 2017 -  Architecture overhaul - Ad serving @ Spotify scale
Qcon London 2017 - Architecture overhaul - Ad serving @ Spotify scale
 
SEMrush: Killer Features
SEMrush: Killer FeaturesSEMrush: Killer Features
SEMrush: Killer Features
 
Marketplace in motion - AdKDD keynote - 2020
Marketplace in motion - AdKDD keynote - 2020 Marketplace in motion - AdKDD keynote - 2020
Marketplace in motion - AdKDD keynote - 2020
 
BOLO2010 Portugal
BOLO2010 PortugalBOLO2010 Portugal
BOLO2010 Portugal
 
Computational Marketing at Groupon - JCSSE 2017
Computational Marketing at Groupon - JCSSE 2017Computational Marketing at Groupon - JCSSE 2017
Computational Marketing at Groupon - JCSSE 2017
 
Flip myfunnel raab v2
Flip myfunnel raab v2Flip myfunnel raab v2
Flip myfunnel raab v2
 
Hardcore PPC Tactics
Hardcore PPC TacticsHardcore PPC Tactics
Hardcore PPC Tactics
 
Ad Extensions as a Lever for Powerful Text Ads
Ad Extensions as a Lever for Powerful Text AdsAd Extensions as a Lever for Powerful Text Ads
Ad Extensions as a Lever for Powerful Text Ads
 
Recsys2016 Tutorial by Xavier and Deepak
Recsys2016 Tutorial by Xavier and DeepakRecsys2016 Tutorial by Xavier and Deepak
Recsys2016 Tutorial by Xavier and Deepak
 
TMKu: Campaign Infrastructure
TMKu: Campaign InfrastructureTMKu: Campaign Infrastructure
TMKu: Campaign Infrastructure
 
Decentralized Advertising Marketplace
Decentralized Advertising MarketplaceDecentralized Advertising Marketplace
Decentralized Advertising Marketplace
 
How to Segment Your Customers and Target Their Needs
How to Segment Your Customers and Target Their NeedsHow to Segment Your Customers and Target Their Needs
How to Segment Your Customers and Target Their Needs
 
Find it! Nail it! Boosting e-commerce search conversions with machine learnin...
Find it! Nail it!Boosting e-commerce search conversions with machine learnin...Find it! Nail it!Boosting e-commerce search conversions with machine learnin...
Find it! Nail it! Boosting e-commerce search conversions with machine learnin...
 
Measurement Plan
Measurement PlanMeasurement Plan
Measurement Plan
 
Google Ad Yield Management (2016 Feb) by Acqua Media
Google Ad Yield Management (2016 Feb) by Acqua MediaGoogle Ad Yield Management (2016 Feb) by Acqua Media
Google Ad Yield Management (2016 Feb) by Acqua Media
 
Google Analytics Premium for Better Data-Driven Decisions With Swapnil Sinha
Google Analytics Premium for Better Data-Driven Decisions With Swapnil SinhaGoogle Analytics Premium for Better Data-Driven Decisions With Swapnil Sinha
Google Analytics Premium for Better Data-Driven Decisions With Swapnil Sinha
 
Intro to Programmatic Advertising with Matt Prohaska from Prohaska Consulting
Intro to Programmatic Advertising with Matt Prohaska from Prohaska ConsultingIntro to Programmatic Advertising with Matt Prohaska from Prohaska Consulting
Intro to Programmatic Advertising with Matt Prohaska from Prohaska Consulting
 
Cross channel attribution overview feb 2010
Cross channel attribution overview feb 2010Cross channel attribution overview feb 2010
Cross channel attribution overview feb 2010
 
Brian Clifton Keynote at Analytics That Excite
Brian Clifton Keynote at Analytics That ExciteBrian Clifton Keynote at Analytics That Excite
Brian Clifton Keynote at Analytics That Excite
 

More from Tiago Albineli Motta

Inteligência Artificial: Da ciência da computação à ciência de dados
Inteligência Artificial: Da ciência da computação à ciência de dadosInteligência Artificial: Da ciência da computação à ciência de dados
Inteligência Artificial: Da ciência da computação à ciência de dadosTiago Albineli Motta
 
Machine Learning no dia a dia do desenvolvedor (Atualizado)
Machine Learning no dia a dia do desenvolvedor (Atualizado)Machine Learning no dia a dia do desenvolvedor (Atualizado)
Machine Learning no dia a dia do desenvolvedor (Atualizado)Tiago Albineli Motta
 
Machine Learning no dia a dia do desenvolvedor
Machine Learning no dia a dia do desenvolvedorMachine Learning no dia a dia do desenvolvedor
Machine Learning no dia a dia do desenvolvedorTiago Albineli Motta
 
Machine Learning e experimentos online para evitar o cancelamento no GloboPlay
Machine Learning e experimentos online para evitar o cancelamento no GloboPlayMachine Learning e experimentos online para evitar o cancelamento no GloboPlay
Machine Learning e experimentos online para evitar o cancelamento no GloboPlayTiago Albineli Motta
 
A ciência de dados por traz de sistemas de recomendação
A ciência de dados por traz de sistemas de recomendaçãoA ciência de dados por traz de sistemas de recomendação
A ciência de dados por traz de sistemas de recomendaçãoTiago Albineli Motta
 
Recomendação de ponta a ponta na Globo.com
Recomendação de ponta a ponta na Globo.comRecomendação de ponta a ponta na Globo.com
Recomendação de ponta a ponta na Globo.comTiago Albineli Motta
 
Testes unitários e de integração: Quando e Porque
Testes unitários e de integração: Quando e PorqueTestes unitários e de integração: Quando e Porque
Testes unitários e de integração: Quando e PorqueTiago Albineli Motta
 

More from Tiago Albineli Motta (18)

Multi Armed Bandit
Multi Armed BanditMulti Armed Bandit
Multi Armed Bandit
 
Inteligência Artificial: Da ciência da computação à ciência de dados
Inteligência Artificial: Da ciência da computação à ciência de dadosInteligência Artificial: Da ciência da computação à ciência de dados
Inteligência Artificial: Da ciência da computação à ciência de dados
 
Machine Learning no dia a dia do desenvolvedor (Atualizado)
Machine Learning no dia a dia do desenvolvedor (Atualizado)Machine Learning no dia a dia do desenvolvedor (Atualizado)
Machine Learning no dia a dia do desenvolvedor (Atualizado)
 
Machine Learning no dia a dia do desenvolvedor
Machine Learning no dia a dia do desenvolvedorMachine Learning no dia a dia do desenvolvedor
Machine Learning no dia a dia do desenvolvedor
 
Experimentation anti patterns
Experimentation anti patternsExperimentation anti patterns
Experimentation anti patterns
 
Machine Learning e experimentos online para evitar o cancelamento no GloboPlay
Machine Learning e experimentos online para evitar o cancelamento no GloboPlayMachine Learning e experimentos online para evitar o cancelamento no GloboPlay
Machine Learning e experimentos online para evitar o cancelamento no GloboPlay
 
A ciência de dados por traz de sistemas de recomendação
A ciência de dados por traz de sistemas de recomendaçãoA ciência de dados por traz de sistemas de recomendação
A ciência de dados por traz de sistemas de recomendação
 
xCLiMF
xCLiMFxCLiMF
xCLiMF
 
Rastros digitais
Rastros digitaisRastros digitais
Rastros digitais
 
Big data
Big dataBig data
Big data
 
Recomendação de ponta a ponta na Globo.com
Recomendação de ponta a ponta na Globo.comRecomendação de ponta a ponta na Globo.com
Recomendação de ponta a ponta na Globo.com
 
Recomendação na Globo.com
Recomendação na Globo.comRecomendação na Globo.com
Recomendação na Globo.com
 
Otimizando seu projeto Rails
Otimizando seu projeto RailsOtimizando seu projeto Rails
Otimizando seu projeto Rails
 
Meta-programacao em python
Meta-programacao em pythonMeta-programacao em python
Meta-programacao em python
 
Testes unitários e de integração: Quando e Porque
Testes unitários e de integração: Quando e PorqueTestes unitários e de integração: Quando e Porque
Testes unitários e de integração: Quando e Porque
 
Redis na Prática
Redis na PráticaRedis na Prática
Redis na Prática
 
Dinamizando Sites Estáticos
Dinamizando Sites EstáticosDinamizando Sites Estáticos
Dinamizando Sites Estáticos
 
Escalando Sites com Nginx
Escalando Sites com NginxEscalando Sites com Nginx
Escalando Sites com Nginx
 

Recently uploaded

Pigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions
 
Unblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesUnblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesSinan KOZAK
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Patryk Bandurski
 
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024BookNet Canada
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationRidwan Fadjar
 
How to convert PDF to text with Nanonets
How to convert PDF to text with NanonetsHow to convert PDF to text with Nanonets
How to convert PDF to text with Nanonetsnaman860154
 
Breaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountBreaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountPuma Security, LLC
 
Artificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning eraArtificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning eraDeakin University
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptxLBM Solutions
 
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024BookNet Canada
 
CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):comworks
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Scott Keck-Warren
 
Understanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitectureUnderstanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitecturePixlogix Infotech
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxMalak Abu Hammad
 
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptxMaking_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptxnull - The Open Security Community
 
Hyderabad Call Girls Khairatabad ✨ 7001305949 ✨ Cheap Price Your Budget
Hyderabad Call Girls Khairatabad ✨ 7001305949 ✨ Cheap Price Your BudgetHyderabad Call Girls Khairatabad ✨ 7001305949 ✨ Cheap Price Your Budget
Hyderabad Call Girls Khairatabad ✨ 7001305949 ✨ Cheap Price Your BudgetEnjoy Anytime
 
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking MenDelhi Call girls
 
Enhancing Worker Digital Experience: A Hands-on Workshop for Partners
Enhancing Worker Digital Experience: A Hands-on Workshop for PartnersEnhancing Worker Digital Experience: A Hands-on Workshop for Partners
Enhancing Worker Digital Experience: A Hands-on Workshop for PartnersThousandEyes
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking MenDelhi Call girls
 

Recently uploaded (20)

Pigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food Manufacturing
 
Unblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesUnblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen Frames
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
 
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
 
The transition to renewables in India.pdf
The transition to renewables in India.pdfThe transition to renewables in India.pdf
The transition to renewables in India.pdf
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 Presentation
 
How to convert PDF to text with Nanonets
How to convert PDF to text with NanonetsHow to convert PDF to text with Nanonets
How to convert PDF to text with Nanonets
 
Breaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountBreaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path Mount
 
Artificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning eraArtificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning era
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptx
 
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
 
CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024
 
Understanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitectureUnderstanding the Laravel MVC Architecture
Understanding the Laravel MVC Architecture
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptx
 
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptxMaking_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
 
Hyderabad Call Girls Khairatabad ✨ 7001305949 ✨ Cheap Price Your Budget
Hyderabad Call Girls Khairatabad ✨ 7001305949 ✨ Cheap Price Your BudgetHyderabad Call Girls Khairatabad ✨ 7001305949 ✨ Cheap Price Your Budget
Hyderabad Call Girls Khairatabad ✨ 7001305949 ✨ Cheap Price Your Budget
 
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
 
Enhancing Worker Digital Experience: A Hands-on Workshop for Partners
Enhancing Worker Digital Experience: A Hands-on Workshop for PartnersEnhancing Worker Digital Experience: A Hands-on Workshop for Partners
Enhancing Worker Digital Experience: A Hands-on Workshop for Partners
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
 

Challenges of Real-Time Recommendations in a Dynamic Marketplace

Editor's Notes

  1. Good morning everybody. Thank you for attending this talk and a special thank to Rodrygo Santos who invited us to talk about our challenges and research for a real-time recommendation in a dynamic marketplace environment. I’m Tiago Motta and I work as Machine Learning Engineer at OLX https://www.flaticon.com
  2. Mainstream recommendation systems often deal items that can be consumed by many users. Examples of unlimited item offerings that we can speak about are movies, digital books, articles and news.
  3. Even when the item consumption are not unlimited, such as products sold online, hotel rooms and food, at least these applications have a way of knowing when the stock runs out, so that the system can stop recommending this item heuristically.
  4. The problem I bring today is completely different. Ignoring some exceptions like ads about job and services We're dealing with unique items that can be consumed by only one buyer And by not doing the intermediation of purchase and sale, we're not sure if the items are not available anymore to stop recommending it.
  5. OLX is the top of mind online classifieds in Brazil, Millions of Brazilians depend on us for living. So we have a huge responsability to improve this product.
  6. We have more than seven million users accessing the platform daily, over five hundred thousand new ads per day. And 15% of contacts between buyers and sellers is through recommended ads, showing that the recommendation system plays a key role in the product.
  7. But this intense stream of new ads prevents us from using traditional collaborative filtering techniques such as Matrix Factorization. That’s because training time for each new batch of ads would make us recommend old ads, slowing down the negotiation and reducing the freshness
  8. That is why our first and current recommendation system is a graph based collaborative filtering. It's calculated through a SQL query, where ads and users are the nodes and implicit ad views feedback are the edges.
  9. The main advantage of this technique is that it allows real-time recommendations without the need for constant and costly retraining This happens because the database is fed in real time by a streaming process, which is then accessed directly by the API.
  10. Despite its simplicity, this technique has ensured us a 6% increase in ad views and a 4% increase in contacts as could be validated in the implemented and runned AB experiments in the past. Even with these great results we have the intuition that it is possible to increase the connectivity and therefore the speed of sales of our sellers
  11. Mainly because the current target metric of our collaborative filtering is the ad visualization and not the action of contacting the seller. However, if we changed this approach adopting contact as the edge of collaborative filtering, we would have the problem of focusing users' attention on ads that already received contact, and that would reduce OLX's democracy.
  12. When this concentration happens it is bad for sellers and buyers. Some sellers have a flood of messages to answer, while others are ignored. And for the buyer is also bad, because he has the object of desire denied because it has already been sold. Thus we began to research an approach that would address both challenges, increase connectivity and increase product democracy.
  13. The idea we are currently validating is to use a classifier to identify which factors in an ad have the greatest influence on stimulating buyer-seller contact given the user's intention when visiting another ad. This would give us a chance to recommend ads that are likely to receive contacts long before it has even been viewed or received any contact. In addition, this would give us the flexibility to heuristically filter ads currently being negotiated, avoiding concentration and giving other ads a chance.
  14. To train this classifier we have created a balanced dataset containing features of both the source and the recommended ad, and as a target label a value informing if the seller was contacted about the recommended ad or not This dataset was created using a pseudo random recommendationdisplayed to a sample users for one week. This randomness was designed to allow us to discover other patterns of consumption that the current recommendation bias would prevent us.
  15. As features of the source and recommended ad, we use the vector representation of the document using title and description generated through doc2vec. The doc2vec model was trained using a sample of 15 million examples of OLX historical ads.
  16. Other important group of features we included was a vector representation of the ad image as the penultimate layer of a pre-trained ResNet18 (eighteen)
  17. In addition, it was natural to imagine that the geographical distance between the ads were a representative factor. As the most granular location information we have from the ad is the neighborhood, we have done a Matrix Factorization between users and the ad neighborhoods where they come in contact with to generate a vector representation of the neighborhood. Look, do not confuse the name “neighborhood” with the term in the algorithm KNN https://web.stanford.edu/~rezab/nips2014workshop/submits/logmat.pdf
  18. Last but not least, the price difference between the ads was included
  19. With all those features in this pseudo random balanced dataset, we use the CatBoost classifier and so far have achieved an accuracy of 75% which is an excellent result given it deals with human behavior
  20. We have not yet been able to validate offline if this technique will do better than our current recommendation. We are currently trying to find a way to validate this fairly and without bias. If anyone have idea of how we can do this, please come talk to us after the talk in the OLX stand But we already have an idea of how to put this algorithm in production.
  21. Once you have this classifier ready, because of our huge ad volume, it will be impracticable to do the prediction in real time on the API So we are thinking of strategies to make this prediction in a background streaming process by calculating the probabilities through the classifier and updating them in real time in a database.
  22. For each new ad, the streaming process will need to: Calculate text and image embeddings and save into the database.
  23. Browse a large sample of ads to calculate the probability of contact using the classifier and save these probabilities to the database. In this moment we are going to have the recommendations for the published item.
  24. Execute the reverse probability calculation, switching the recommended ad as source ad to update recommendations for old items in the database
  25. That way the API would only need to execute an optimized SQL query while maintaining our excellent performance.
  26. As future research and improvements we are going to experiment the embedding technique adopted by Gabriel Moreiras’s in his excelent paper about Chamaleon. And as an alternative on the classification we are thinking about doing the matrix factorization using the items features as Kula described in his paper from recsys 2015 and implemented in lightFM. https://github.com/lyst/lightfm https://dl.acm.org/citation.cfm?doid=3270323.3270328
  27. We hope this item-item content-based classification technique will allow us to have the flexibility of a wide range of new offerings with many different types of filtering. However, only offline and online validation of this technique will show us if it can improve OLX connectivity and democracy over the previous recommendation. We hope to have this result at RecSys 2020 in Rio de Janeiro, we are going to wait for you there.